Catalog / Business processes / Marketing / Inquiry-to-Owner / Maintain Lead Scoring Model
reference process · marketing · 18 activities · 13 on the rosterMaintain Lead Scoring Model
The scoring model behind inbound lead routing stops matching what leads actually do, or a new fact about a lead becomes available that the model does not read. This process is how that model changes. An analytics agent reads how the scores of the last period compare with what those leads went on to do, a forecaster fits two or three candidate models on the organization's own record, and each candidate is run over leads that have already closed and then run on live leads beside the model in production. The people who work the leads sit in the room where one candidate is chosen. The run ends with the chosen model issued as a numbered version, scoring every lead, and watched through its first weeks against what those leads do. Lead routing keeps running on the current version the whole time this process is open.
The walk
The document
One run
Adoption
The activities
What happens in a run
18 activities from scores stop predicting, or a new signal appears to the recalibrated model is in production, 2 of them gates a person has to sign. Drag the diagram to move along it.
Maintain Lead Scoring Model scores stop predicting, or a new signal appears → the recalibrated model is in production
a source is ruled out the signal arrives too late no candidate beats the model in use a territory queue would be flooded the room cannot agree it behaves differently on live leads the new version scores worse scores stop predicting, or a new signal appears 1 Take in the Drift or the … work out what the model must fix and by when 2 Read How the Model Has Sc… scores set against what those leads went on to do 3 Read the Exception Queue the leads no rule could settle, and how they settled 4 Hear from the People Who … where the scores were wrong, in their own words 5 Gather the Candidate Sign… facts that might predict a win, and their sources 6 Clear the New Signals the terms on each source, and what may be scored on a person signs · never an agent 7 Test the Signal Coverage how often each signal is filled in, and how early 8 Fit the Candidate Models two or three candidates, each with its assumptions 9 Back-Test the Candidates each candidate run over the leads already closed 10 Check What Each Does to R… how the tiers and the territory queues would shift 11 Choose the Model one candidate chosen, with the reasons said out loud a person signs · never an agent 12 Run It Beside the Live Mo… both score live leads, only the current one routes 13 Issue the Model as a Vers… the model numbered, and the agents that read it told 14 Hand over the Model each receiving agent says what it changes 15 Put It into Production the new version starts scoring every lead 16 Watch the First Weeks the new scores set against what the leads did 17 Set the Review Conditions the review date, and what brings the model back 18 Record What Was Learned what to repeat, what to avoid the recalibrated model is in production The roster this process needs
Hover a name to see the activities it holds. A dashed one is a person, and stays one.
The document
The document, with the blanks marked
Everything in amber is yours to fill in: who owns it, when it takes effect, which platforms, which numbers, and who holds each activity. Everything else is the process, and it is the same wherever it is run.
PROCESS: lead scoring model id: <team> /lead-scoring-model v1
from: ref/mkt/lead-scoring-model v1
owner: <who> effective: <date>
trigger: the accuracy report on <your lead routing process> shows the
scores have drifted, or a new signal about a lead becomes
available, or the review date on the model in production
arrives
watch: record=<accuracy report>
system=<your lead routing process>
change=<the report shows the scores have drifted>
or watch: record=<candidate signal>
system=<the source the signal is read from>
change=<a new fact about a lead becomes available>
or watch: record=<model in production>
system=<your model store>
change=<the review date on the model arrives>
concurrency: one run at a time for one scoring model, because two runs
would put two models into production. Runs against
different models overlap, and lead routing keeps scoring
on the current version throughout
goal: one recalibrated model in production as a numbered version, with
the back-test, the routing effect and the shadow run behind it,
and every agent that reads a score told what changed
phases:
take-in-trigger - convenes briefing: what the model has to fix, who
decides it, what evidence each agent owes, and the
date the answer is due
owner: decision-coordinator after: trigger
automation: <level>
read-scoring - human: the scores of the last <period> set against
what those leads went on to do, split by tier
owner: analytics after: take-in-trigger
by: <days> automation: <level>
read-exceptions - system: the leads that could not be scored or
routed, and what a person decided about each one
owner: lead-scorer after: take-in-trigger
automation: <level>
first-hand - human: what the people who work the leads say
about the scores those leads arrived with
owner: researcher after: take-in-trigger
by: <days> automation: <level>
gather-signals - human: the facts that might predict a win, each
with the source it would be read from
owner: forecaster
after: read-scoring + read-exceptions + first-hand
by: <days> automation: <level>
clear-signals - convenes approval: legal signs that each new
source may be used to score a person, and the
consent record says which contacts it covers
owner: <your legal role> after: gather-signals
by: <days> automation: never
test-coverage - system: how often each signal is already filled
in, what the rest cost to fill, and which ones
arrive after the lead has been routed
owner: lead-scorer after: gather-signals
automation: <level>
fit-candidates - runs build-by-talent: two or three candidate
models, each carrying the data it was fit on and
the assumptions it rests on
owner: forecaster
after: clear-signals + test-coverage
by: <days> automation: <level>
back-test - system: each candidate run over the leads that
have already closed, to see what it would have
scored them and what those leads did
owner: forecaster after: fit-candidates
by: <days> automation: <level>
routing-effect - system: how many leads land in each tier and each
territory queue under each candidate
owner: lead-scorer after: fit-candidates
automation: <level>
choose-model - convenes bake-off: two or three candidates, one
chosen, with the reasons said out loud
owner: decision-coordinator
after: back-test + routing-effect
by: <days> automation: never
shadow-run - system: both models score every live lead for
<how long> , and only the version in production
routes anything
owner: lead-scorer after: choose-model
by: <days> automation: <level>
issue-version - human: the chosen model issued as a numbered
version, with the version it replaces left readable
owner: standards-keeper after: shadow-run
automation: <level>
hand-over - runs decide-and-announce: every agent that reads a
score hears the new version at once and says what
it changes because of it
owner: decision-coordinator after: issue-version
by: <days> automation: <level>
cut-over - system: the new version starts scoring every lead,
and the version it replaced stays readable
owner: lead-scorer after: hand-over
automation: <level>
watch-new-model - human: the first <how long> of scores set against
what those leads went on to do
owner: analytics after: cut-over
by: <weeks> automation: <level>
set-review - human: the date the model comes back up, and the
drift that brings it back sooner
owner: decision-coordinator
after: watch-new-model automation: <level>
record-learnings - convenes debrief: what to repeat, what to avoid
owner: decision-coordinator after: set-review
automation: <level>
run-scoped:
outstanding - runs roll-call owner: decision-coordinator
every: <cadence>
from: take-in-trigger until: hand-over
accuracy - runs collect-and-report owner: analytics
every: <cadence>
from: cut-over until: run close
handoffs:
take-in-trigger -> read-scoring / read-exceptions / first-hand
[what-to-fix]: what the model has to fix, what each agent owes, and
the date it is owed. Three agents read one record
read-scoring -> gather-signals [scores-versus-outcomes]: the scores of
the period set beside what each lead went on to do, each score
naming the model version it was computed under
read-exceptions -> gather-signals [exception-decisions]: the leads no
rule could settle, what the person decided about each one, and the
field that was missing
first-hand -> gather-signals [first-hand-accounts]: what the people
who work the leads said, attributed to the person who said it
gather-signals -> clear-signals [candidate-signals]: each candidate
signal, the source it would be read from, and the terms that source
came with
gather-signals -> test-coverage [signals-to-test]: the same signals,
in the form they would be read in. Two branches read one record
clear-signals -> fit-candidates [cleared-signals]: the signals the
model may be fit on, and which contacts the consent record covers
for each one
test-coverage -> fit-candidates [coverage-findings]: how often each
signal is present, what the rest cost to fill, and which ones arrive
after the lead has been routed
fit-candidates -> back-test [candidate-models]: each candidate at a
version, with the data it was fit on and the assumptions it rests on
fit-candidates -> routing-effect [candidates-to-route]: the same
candidates. Two branches read one record
back-test -> choose-model [back-test-results]: what each candidate
would have scored the closed leads, and what those leads actually
did
routing-effect -> choose-model [routing-shift]: how many leads land in
each tier and each territory queue under each candidate, set beside
what the model in production does now
choose-model -> shadow-run [chosen-model]: the chosen candidate, why
it won over the others, and everyone who disagreed and on what
shadow-run -> issue-version [shadow-scores]: both scores for every
lead of the shadow period, and every lead the two models would have
routed to different owners
issue-version -> hand-over [versioned-model]: the model at its number,
and what changed from the version it replaces
hand-over -> cut-over [adoption-answers]: what each receiving agent
said it would change, and the date it said it by
cut-over -> watch-new-model [cut-over-record]: the date the new
version started scoring, and the last lead the version it replaced
scored
watch-new-model -> set-review [first-weeks-scores]: the first weeks of
scores set against what those leads did
set-review -> record-learnings [review-conditions]: the review date,
and the drift that brings the model back sooner
deviations:
clear-signals -> gather-signals [source-ruled-out]: legal rules out a
source, so the signals are gathered again without it
test-coverage -> gather-signals [signal-too-late]: a signal arrives
after the lead has been routed, so the signals are gathered again
back-test -> fit-candidates [no-candidate-beats]: no candidate beats
the model already in use, so the candidates are fit again on a
different set of signals
routing-effect -> fit-candidates [queue-flooded]: a candidate would
flood a territory queue, so the candidates are fit again or their
thresholds are changed
choose-model -> fit-candidates [room-cannot-agree]: the room cannot
agree on one candidate, so the candidates are fit again against what
the disagreement showed
shadow-run -> fit-candidates [shadow-disagrees]: the chosen candidate
behaves differently on live leads than it did on the closed ones, so
the candidates are fit again with the disputed leads attached
watch-new-model -> choose-model [scores-worse]: the new version scores
worse than the one it replaced, so the choice is made again and
whichever way it goes runs through the shadow, the version and the
handover
bindings:
roster: <who holds each role - agents claiming the abstract agents
above, and named people for the sales side, leadership and
legal>
systems: the CRM (read), enrichment sources (read),
analytics platforms (read),
the customer and revenue record (read),
the consent record (read), the model store (write),
the standards store (write), the decision record (write),
the review schedule (write),
the notification channel (write)
data: the model in production at its version, the territory map
at its version, <your lead record> over <what period> ,
the terms attached to every source the model reads
policy:
every score records the model version and the inputs it was computed
from
a model in production is never edited in place. A change goes in as
a new numbered version, and the version it replaces stays readable
no candidate reaches the room without a back-test over the leads
that have already closed
a signal that cannot be read before the lead is routed is marked as
such, and <who> decides whether the model may score on it
a rate the model cannot source is reported as unsourced, and no
estimate is put in its place
no new signal is scored on until legal has signed that its source
may be used and the consent record has been applied to it
the model that goes into production is chosen by named people, and
the choice is never delegated to an agent
the people who work the leads sit in the room where the model is
chosen, and a model chosen without them is not issued
the run does not close until every agent that reads a score has said
what it changes, and an agent that changes nothing says why
measures:
cycle time: <target> from the trigger to the new version scoring
accuracy: <share> of the leads scored in the top tier that went on
to <what counts as a win> , read on the review date
coverage: <share> of the signals the model reads that are filled in
before a lead is routed
quality gate: no version reaches production without its back-test,
its routing effect and its shadow run
Copy
Take it somewhere
Use this process in LangGraph
close
Paste this into an assistant that can read the web, such as Claude, ChatGPT or Cursor. It reads the specification and the current LangGraph documentation, then writes two files: the graph, and a note on what did not survive the translation. Read the note first. What a runtime cannot express is the part worth arguing about, and this process is a draft to argue with.
copy the prompt
346 lines · the document is inside it, so nothing else is needed
Convert the business process below into a runnable LangGraph graph:
one Python file with a TypedDict state, a StateGraph, nodes, edges,
conditional edges and a checkpointer.
The document is a reference process written to the Agent Processes
specification. Read the specification before you start, because it defines
terms that look ordinary and are not:
https://agentcatalog.com/spec/agent-processes
Sections 6 (the phase graph), 6.5.1 (exception edges), 6.7.1 (deviations),
7 (automation) and 8 (handoffs) are the ones this conversion turns on.
Then read the current documentation for the primitives you will need, rather
than relying on what you remember of the API:
https://docs.langchain.com/oss/python/langgraph/interrupts
interrupt() and Command(resume=), which is how a gate stops a run
https://reference.langchain.com/python/langgraph/graph/state/StateGraph
StateGraph, add_edge, add_conditional_edges, defer
WHAT THE DOCUMENT ASKS FOR
These hold wherever the process lands, and they matter more than style.
1. Each phase under `phases:` becomes one step, and keeps its name.
2. `after:` gives the edges. `after: a + b` is a join and waits for BOTH.
Reading it as "either" is the defect the specification calls out by name.
3. Every handoff carries a key in square brackets. Each key becomes one field
on the run's state, named exactly as the key with hyphens turned into
underscores, and the sentence beside it becomes that field's comment. The key
is the stable name; the sentence is prose that may be rewritten.
4. A phase MUST NOT begin before its inbound handoff exists. Where that is
checkable, check it in the step rather than assuming it.
5. `automation: never` is a gate a person signs. The run stops there and does
not continue until a person's decision comes back. Do not turn one into a
notification, a log line, or an automatic transition, whatever the queue
looks like.
6. Each line under `deviations:` is a backward or sideways edge, returning to
the phase named on the right. The key in brackets names it, and that name
belongs in the code.
7. A phase whose `after:` reads like "X or Y, whichever could not finish" is an
exception edge: it is entered when those phases FAIL, not when they succeed.
Do not wire it as an ordinary successor.
8. Anything in angle brackets is a blank the adopting organization fills in.
Leave each one as a named constant at the top of the file with a TODO. Do not
invent a value, a threshold or a date.
9. Record the document's `from:` line at the top of the file, so it says which
reference process and which version it was generated from.
10. Run-scoped lines under `run-scoped:` are work that runs alongside the whole
process rather than at one point in it, and a run may not close while one is
unfinished. Say in the code what you did about them, including if the answer
is that the runtime has nowhere to put them.
HOW THAT LOOKS IN LANGGRAPH
11. A phase is a node added with `add_node`, under the phase's own name.
12. The state is a TypedDict. Each handoff key is one field on it.
13. A join is the trap. `add_edge(["a", "b"], "c")` looks right and releases
once: when a backward edge re-enters ONE arm, the joined node never runs
again, and the run ends early reporting success rather than raising. Mark the
joined node `defer=True` and re-check inside it that both inbound handoffs
exist.
14. A gate is `interrupt()` inside the node, resumed with `Command(resume=...)`.
The platform lets anything at all call resume, so require the resumed value to
name a person and a date and refuse anything else. Say in the fidelity note
that this proves only that whoever resumed typed a name, because
`Command(resume=True)` from a scheduled job is indistinguishable from a person
signing.
15. A deviation is `add_conditional_edges` with a routing function named after
the key in brackets.
16. Pass a durable checkpointer rather than taking the in-memory default. The
gates wait days, and the default loses every paused run on restart.
17. Leave every phase body unimplemented, raising until somebody registers an
implementation. The automation level is a blank, so writing a body would
answer on the adopter's behalf whether an agent may do that work.
Produce a second file alongside it, `FIDELITY.md`, and treat it as the more
important of the two. The code is for whoever builds this. The fidelity note
is for whoever has to decide whether this platform suits the process at all,
and that is usually a different person who will never read the code.
It has three parts.
**What came across.** Briefly: how many phases became steps, how many handoff
keys became state fields, which gates stop the run, which deviations became
edges. Counts and names, not reassurance.
**What did not, and what was done instead.** One entry per gap. For each one,
say what the document requires, what the platform can actually express, what
you did in its place, and what breaks if somebody later removes your
workaround. This last part matters most: a workaround nobody understands is a
workaround somebody deletes.
**What a person still has to decide.** The blanks are not a translation
failure, they are the point of a reference process, so list what has to be
filled in before this could run against anything real, and say which of those
choices the platform constrains.
Write it in plain English for somebody who has not read the specification, and
do not soften it. A translation of a reference process is a draft to argue
with, not a build artifact, and the honest account of what was lost is the most
useful thing you will produce.
Here is the process document.
```
PROCESS: lead scoring model id: <team>/lead-scoring-model v1
from: ref/mkt/lead-scoring-model v1
owner: <who> effective: <date>
trigger: the accuracy report on <your lead routing process> shows the
scores have drifted, or a new signal about a lead becomes
available, or the review date on the model in production
arrives
watch: record=<accuracy report>
system=<your lead routing process>
change=<the report shows the scores have drifted>
or watch: record=<candidate signal>
system=<the source the signal is read from>
change=<a new fact about a lead becomes available>
or watch: record=<model in production>
system=<your model store>
change=<the review date on the model arrives>
concurrency: one run at a time for one scoring model, because two runs
would put two models into production. Runs against
different models overlap, and lead routing keeps scoring
on the current version throughout
goal: one recalibrated model in production as a numbered version, with
the back-test, the routing effect and the shadow run behind it,
and every agent that reads a score told what changed
phases:
take-in-trigger - convenes briefing: what the model has to fix, who
decides it, what evidence each agent owes, and the
date the answer is due
owner: decision-coordinator after: trigger
automation: <level>
read-scoring - human: the scores of the last <period> set against
what those leads went on to do, split by tier
owner: analytics after: take-in-trigger
by: <days> automation: <level>
read-exceptions - system: the leads that could not be scored or
routed, and what a person decided about each one
owner: lead-scorer after: take-in-trigger
automation: <level>
first-hand - human: what the people who work the leads say
about the scores those leads arrived with
owner: researcher after: take-in-trigger
by: <days> automation: <level>
gather-signals - human: the facts that might predict a win, each
with the source it would be read from
owner: forecaster
after: read-scoring + read-exceptions + first-hand
by: <days> automation: <level>
clear-signals - convenes approval: legal signs that each new
source may be used to score a person, and the
consent record says which contacts it covers
owner: <your legal role> after: gather-signals
by: <days> automation: never
test-coverage - system: how often each signal is already filled
in, what the rest cost to fill, and which ones
arrive after the lead has been routed
owner: lead-scorer after: gather-signals
automation: <level>
fit-candidates - runs build-by-talent: two or three candidate
models, each carrying the data it was fit on and
the assumptions it rests on
owner: forecaster
after: clear-signals + test-coverage
by: <days> automation: <level>
back-test - system: each candidate run over the leads that
have already closed, to see what it would have
scored them and what those leads did
owner: forecaster after: fit-candidates
by: <days> automation: <level>
routing-effect - system: how many leads land in each tier and each
territory queue under each candidate
owner: lead-scorer after: fit-candidates
automation: <level>
choose-model - convenes bake-off: two or three candidates, one
chosen, with the reasons said out loud
owner: decision-coordinator
after: back-test + routing-effect
by: <days> automation: never
shadow-run - system: both models score every live lead for
<how long>, and only the version in production
routes anything
owner: lead-scorer after: choose-model
by: <days> automation: <level>
issue-version - human: the chosen model issued as a numbered
version, with the version it replaces left readable
owner: standards-keeper after: shadow-run
automation: <level>
hand-over - runs decide-and-announce: every agent that reads a
score hears the new version at once and says what
it changes because of it
owner: decision-coordinator after: issue-version
by: <days> automation: <level>
cut-over - system: the new version starts scoring every lead,
and the version it replaced stays readable
owner: lead-scorer after: hand-over
automation: <level>
watch-new-model - human: the first <how long> of scores set against
what those leads went on to do
owner: analytics after: cut-over
by: <weeks> automation: <level>
set-review - human: the date the model comes back up, and the
drift that brings it back sooner
owner: decision-coordinator
after: watch-new-model automation: <level>
record-learnings - convenes debrief: what to repeat, what to avoid
owner: decision-coordinator after: set-review
automation: <level>
run-scoped:
outstanding - runs roll-call owner: decision-coordinator
every: <cadence>
from: take-in-trigger until: hand-over
accuracy - runs collect-and-report owner: analytics
every: <cadence>
from: cut-over until: run close
handoffs:
take-in-trigger -> read-scoring / read-exceptions / first-hand
[what-to-fix]: what the model has to fix, what each agent owes, and
the date it is owed. Three agents read one record
read-scoring -> gather-signals [scores-versus-outcomes]: the scores of
the period set beside what each lead went on to do, each score
naming the model version it was computed under
read-exceptions -> gather-signals [exception-decisions]: the leads no
rule could settle, what the person decided about each one, and the
field that was missing
first-hand -> gather-signals [first-hand-accounts]: what the people
who work the leads said, attributed to the person who said it
gather-signals -> clear-signals [candidate-signals]: each candidate
signal, the source it would be read from, and the terms that source
came with
gather-signals -> test-coverage [signals-to-test]: the same signals,
in the form they would be read in. Two branches read one record
clear-signals -> fit-candidates [cleared-signals]: the signals the
model may be fit on, and which contacts the consent record covers
for each one
test-coverage -> fit-candidates [coverage-findings]: how often each
signal is present, what the rest cost to fill, and which ones arrive
after the lead has been routed
fit-candidates -> back-test [candidate-models]: each candidate at a
version, with the data it was fit on and the assumptions it rests on
fit-candidates -> routing-effect [candidates-to-route]: the same
candidates. Two branches read one record
back-test -> choose-model [back-test-results]: what each candidate
would have scored the closed leads, and what those leads actually
did
routing-effect -> choose-model [routing-shift]: how many leads land in
each tier and each territory queue under each candidate, set beside
what the model in production does now
choose-model -> shadow-run [chosen-model]: the chosen candidate, why
it won over the others, and everyone who disagreed and on what
shadow-run -> issue-version [shadow-scores]: both scores for every
lead of the shadow period, and every lead the two models would have
routed to different owners
issue-version -> hand-over [versioned-model]: the model at its number,
and what changed from the version it replaces
hand-over -> cut-over [adoption-answers]: what each receiving agent
said it would change, and the date it said it by
cut-over -> watch-new-model [cut-over-record]: the date the new
version started scoring, and the last lead the version it replaced
scored
watch-new-model -> set-review [first-weeks-scores]: the first weeks of
scores set against what those leads did
set-review -> record-learnings [review-conditions]: the review date,
and the drift that brings the model back sooner
deviations:
clear-signals -> gather-signals [source-ruled-out]: legal rules out a
source, so the signals are gathered again without it
test-coverage -> gather-signals [signal-too-late]: a signal arrives
after the lead has been routed, so the signals are gathered again
back-test -> fit-candidates [no-candidate-beats]: no candidate beats
the model already in use, so the candidates are fit again on a
different set of signals
routing-effect -> fit-candidates [queue-flooded]: a candidate would
flood a territory queue, so the candidates are fit again or their
thresholds are changed
choose-model -> fit-candidates [room-cannot-agree]: the room cannot
agree on one candidate, so the candidates are fit again against what
the disagreement showed
shadow-run -> fit-candidates [shadow-disagrees]: the chosen candidate
behaves differently on live leads than it did on the closed ones, so
the candidates are fit again with the disputed leads attached
watch-new-model -> choose-model [scores-worse]: the new version scores
worse than the one it replaced, so the choice is made again and
whichever way it goes runs through the shadow, the version and the
handover
bindings:
roster: <who holds each role - agents claiming the abstract agents
above, and named people for the sales side, leadership and
legal>
systems: the CRM (read), enrichment sources (read),
analytics platforms (read),
the customer and revenue record (read),
the consent record (read), the model store (write),
the standards store (write), the decision record (write),
the review schedule (write),
the notification channel (write)
data: the model in production at its version, the territory map
at its version, <your lead record> over <what period>,
the terms attached to every source the model reads
policy:
every score records the model version and the inputs it was computed
from
a model in production is never edited in place. A change goes in as
a new numbered version, and the version it replaces stays readable
no candidate reaches the room without a back-test over the leads
that have already closed
a signal that cannot be read before the lead is routed is marked as
such, and <who> decides whether the model may score on it
a rate the model cannot source is reported as unsourced, and no
estimate is put in its place
no new signal is scored on until legal has signed that its source
may be used and the consent record has been applied to it
the model that goes into production is chosen by named people, and
the choice is never delegated to an agent
the people who work the leads sit in the room where the model is
chosen, and a model chosen without them is not issued
the run does not close until every agent that reads a score has said
what it changes, and an agent that changes nothing says why
measures:
cycle time: <target> from the trigger to the new version scoring
accuracy: <share> of the leads scored in the top tier that went on
to <what counts as a win>, read on the review date
coverage: <share> of the signals the model reads that are filled in
before a lead is routed
quality gate: no version reaches production without its back-test,
its routing effect and its shadow run
```
Take it somewhere
Use this process in Microsoft Agent Framework
close
Paste this into an assistant that can read the web, such as Claude, ChatGPT or Cursor. It reads the specification and the current Agent Framework documentation, then writes two files: the workflow, and a note on what did not survive the translation. Read the note first. What a runtime cannot express is the part worth arguing about, and this process is a draft to argue with.
copy the prompt
401 lines · the document is inside it, so nothing else is needed
Convert the business process below into a runnable Microsoft Agent Framework workflow:
one Python file with Executor classes, a WorkflowBuilder, typed edges,
request_info gates and durable checkpoint storage.
The document is a reference process written to the Agent Processes
specification. Read the specification before you start, because it defines
terms that look ordinary and are not:
https://agentcatalog.com/spec/agent-processes
Sections 6 (the phase graph), 6.5.1 (exception edges), 6.7.1 (deviations),
7 (automation) and 8 (handoffs) are the ones this conversion turns on.
Then read the current documentation for the primitives you will need, rather
than relying on what you remember of the API:
https://learn.microsoft.com/en-us/agent-framework/workflows/human-in-the-loop
ctx.request_info, @response_handler, and answering a parked run later
https://learn.microsoft.com/en-us/agent-framework/workflows/checkpoints
what a checkpoint holds, and allowed_checkpoint_types
https://learn.microsoft.com/en-us/agent-framework/concepts/workflows/edges
add_edge with condition, add_fan_in_edges, add_switch_case_edge_group
https://learn.microsoft.com/en-us/agent-framework/concepts/workflows/state
ctx.set_state and ctx.get_state as they actually are today
WHAT THE DOCUMENT ASKS FOR
These hold wherever the process lands, and they matter more than style.
1. Each phase under `phases:` becomes one step, and keeps its name.
2. `after:` gives the edges. `after: a + b` is a join and waits for BOTH.
Reading it as "either" is the defect the specification calls out by name.
3. Every handoff carries a key in square brackets. Each key becomes one field
on the run's state, named exactly as the key with hyphens turned into
underscores, and the sentence beside it becomes that field's comment. The key
is the stable name; the sentence is prose that may be rewritten.
4. A phase MUST NOT begin before its inbound handoff exists. Where that is
checkable, check it in the step rather than assuming it.
5. `automation: never` is a gate a person signs. The run stops there and does
not continue until a person's decision comes back. Do not turn one into a
notification, a log line, or an automatic transition, whatever the queue
looks like.
6. Each line under `deviations:` is a backward or sideways edge, returning to
the phase named on the right. The key in brackets names it, and that name
belongs in the code.
7. A phase whose `after:` reads like "X or Y, whichever could not finish" is an
exception edge: it is entered when those phases FAIL, not when they succeed.
Do not wire it as an ordinary successor.
8. Anything in angle brackets is a blank the adopting organization fills in.
Leave each one as a named constant at the top of the file with a TODO. Do not
invent a value, a threshold or a date.
9. Record the document's `from:` line at the top of the file, so it says which
reference process and which version it was generated from.
10. Run-scoped lines under `run-scoped:` are work that runs alongside the whole
process rather than at one point in it, and a run may not close while one is
unfinished. Say in the code what you did about them, including if the answer
is that the runtime has nowhere to put them.
HOW THAT LOOKS IN MICROSOFT AGENT FRAMEWORK
11. Write for Python, and read those pages before writing a line. This API has
moved: `set_shared_state`, `RequestInfoExecutor`, `RequestInfoMessage` and
`send_responses_streaming` are all in the training data and none of them exist
any more. The .NET workflow API differs in kind rather than in spelling, so a
file written for one does not port by renaming.
12. A phase is a class deriving from `Executor` whose `super().__init__(id=...)`
takes the phase name verbatim. The id is not cosmetic: a checkpoint stores a
signature over the topology and the executor ids, so an id built from a run, a
timestamp or a counter cannot be resumed into.
13. Give each phase a `@handler` method and move work on with
`await ctx.send_message(...)`. A handler that returns without sending is a dead
end: the branch stops, the run converges, and it reports success. So a phase
you are deliberately leaving unimplemented must still send a placeholder
onward. Do not stub with `raise NotImplementedError`, which fails the run
instead of leaving it runnable.
14. `after: a` is `builder.add_edge(a, b)`. `after: a + b` is
`builder.add_fan_in_edges([a, b], target)`, and the target's handler must be
annotated `list[T]`, because a fan-in delivers one aggregated list rather than
the separate messages. A handler typed for the single value is dropped as a
mismatch with nothing raised.
15. The join is the trap here, and it is the opposite of LangGraph's. The
barrier re-arms: it clears its buffer when it fires and then demands a fresh
message from every source. So a deviation that re-enters ONE arm parks the
run forever waiting for an arm that will not run again, and the workflow ends
IDLE reporting success. Wherever a deviation re-enters one arm of a join,
replace the barrier with an ordinary edge from each arm into a small executor
that records each arrival with `ctx.set_state` and only forwards when every
expected key is present, and say in a comment that putting `add_fan_in_edges`
back reintroduces the stall.
16. A phase that is both a join target and a deviation target needs two
handlers, one annotated `list[T]` for the barrier and one annotated `T` for the
backward message. Write only the list handler and every backward edge into it
is discarded as a type mismatch, silently.
17. `automation: never` is `await ctx.request_info(request_data=...,
response_type=...)` inside the phase, answered by a `@response_handler` on the
same executor whose annotations match those exact types. The run parks at
`IDLE_WITH_PENDING_REQUESTS` and the host answers with
`workflow.run(stream=True, responses={request_id: value})`. If no handler
matches the pair, the framework logs a warning and parks anyway, so the gate
reads as working right up until somebody asks why the approval did not take.
18. A gate is only a gate if the wait survives a restart, so pass
`checkpoint_storage=FileCheckpointStorage(...)` to the builder. Checkpointing
is off by default and `InMemoryCheckpointStorage` reads as configured while
persisting nothing. Register every handoff payload type in
`allowed_checkpoint_types`, or the first restore raises. Anything an executor
keeps as an instance attribute is absent after a restore unless you export it
from `on_checkpoint_save` and read it back in `on_checkpoint_restore`, and it
comes back empty rather than missing.
19. Each line under `deviations:` is `builder.add_edge(source, earlier,
condition=fn)` with `fn` named after the key. Cycles are legal and unchecked,
but raise `max_iterations` well above its default of 100, because several live
cycles will exhaust a budget sized for a straight line and fail with a message
about convergence that reads like a broken graph. Keep the ordinary forward
edge unconditional and add each deviation beside it: a condition that returns
false is dropped with no event, so a forward path expressed as a condition
dies silently on every normal run, which is most of them.
20. Handoff values go in `ctx.set_state(key, value)` and come back from
`ctx.get_state(key)`, untyped and unchecked. A write is visible to its writer
at once and to everyone else only in the next superstep, and two writers of one
key in a superstep keep the last write. Never read a key in the same superstep
another phase wrote it.
21. There are no timers, no deadlines and no scheduled wakes. Nothing in
`run-scoped:` becomes an executor and `by:` has no expression at all, so write
them as comments naming where they start and stop, and say plainly in the
fidelity note that a run can close over an unfinished run-scoped line. Do not
fake a deadline with a sleep inside a handler, which blocks the whole superstep
barrier, and never let an expiring wait release a gate.
Produce a second file alongside it, `FIDELITY.md`, and treat it as the more
important of the two. The code is for whoever builds this. The fidelity note
is for whoever has to decide whether this platform suits the process at all,
and that is usually a different person who will never read the code.
It has three parts.
**What came across.** Briefly: how many phases became steps, how many handoff
keys became state fields, which gates stop the run, which deviations became
edges. Counts and names, not reassurance.
**What did not, and what was done instead.** One entry per gap. For each one,
say what the document requires, what the platform can actually express, what
you did in its place, and what breaks if somebody later removes your
workaround. This last part matters most: a workaround nobody understands is a
workaround somebody deletes.
**What a person still has to decide.** The blanks are not a translation
failure, they are the point of a reference process, so list what has to be
filled in before this could run against anything real, and say which of those
choices the platform constrains.
Write it in plain English for somebody who has not read the specification, and
do not soften it. A translation of a reference process is a draft to argue
with, not a build artifact, and the honest account of what was lost is the most
useful thing you will produce.
Here is the process document.
```
PROCESS: lead scoring model id: <team>/lead-scoring-model v1
from: ref/mkt/lead-scoring-model v1
owner: <who> effective: <date>
trigger: the accuracy report on <your lead routing process> shows the
scores have drifted, or a new signal about a lead becomes
available, or the review date on the model in production
arrives
watch: record=<accuracy report>
system=<your lead routing process>
change=<the report shows the scores have drifted>
or watch: record=<candidate signal>
system=<the source the signal is read from>
change=<a new fact about a lead becomes available>
or watch: record=<model in production>
system=<your model store>
change=<the review date on the model arrives>
concurrency: one run at a time for one scoring model, because two runs
would put two models into production. Runs against
different models overlap, and lead routing keeps scoring
on the current version throughout
goal: one recalibrated model in production as a numbered version, with
the back-test, the routing effect and the shadow run behind it,
and every agent that reads a score told what changed
phases:
take-in-trigger - convenes briefing: what the model has to fix, who
decides it, what evidence each agent owes, and the
date the answer is due
owner: decision-coordinator after: trigger
automation: <level>
read-scoring - human: the scores of the last <period> set against
what those leads went on to do, split by tier
owner: analytics after: take-in-trigger
by: <days> automation: <level>
read-exceptions - system: the leads that could not be scored or
routed, and what a person decided about each one
owner: lead-scorer after: take-in-trigger
automation: <level>
first-hand - human: what the people who work the leads say
about the scores those leads arrived with
owner: researcher after: take-in-trigger
by: <days> automation: <level>
gather-signals - human: the facts that might predict a win, each
with the source it would be read from
owner: forecaster
after: read-scoring + read-exceptions + first-hand
by: <days> automation: <level>
clear-signals - convenes approval: legal signs that each new
source may be used to score a person, and the
consent record says which contacts it covers
owner: <your legal role> after: gather-signals
by: <days> automation: never
test-coverage - system: how often each signal is already filled
in, what the rest cost to fill, and which ones
arrive after the lead has been routed
owner: lead-scorer after: gather-signals
automation: <level>
fit-candidates - runs build-by-talent: two or three candidate
models, each carrying the data it was fit on and
the assumptions it rests on
owner: forecaster
after: clear-signals + test-coverage
by: <days> automation: <level>
back-test - system: each candidate run over the leads that
have already closed, to see what it would have
scored them and what those leads did
owner: forecaster after: fit-candidates
by: <days> automation: <level>
routing-effect - system: how many leads land in each tier and each
territory queue under each candidate
owner: lead-scorer after: fit-candidates
automation: <level>
choose-model - convenes bake-off: two or three candidates, one
chosen, with the reasons said out loud
owner: decision-coordinator
after: back-test + routing-effect
by: <days> automation: never
shadow-run - system: both models score every live lead for
<how long>, and only the version in production
routes anything
owner: lead-scorer after: choose-model
by: <days> automation: <level>
issue-version - human: the chosen model issued as a numbered
version, with the version it replaces left readable
owner: standards-keeper after: shadow-run
automation: <level>
hand-over - runs decide-and-announce: every agent that reads a
score hears the new version at once and says what
it changes because of it
owner: decision-coordinator after: issue-version
by: <days> automation: <level>
cut-over - system: the new version starts scoring every lead,
and the version it replaced stays readable
owner: lead-scorer after: hand-over
automation: <level>
watch-new-model - human: the first <how long> of scores set against
what those leads went on to do
owner: analytics after: cut-over
by: <weeks> automation: <level>
set-review - human: the date the model comes back up, and the
drift that brings it back sooner
owner: decision-coordinator
after: watch-new-model automation: <level>
record-learnings - convenes debrief: what to repeat, what to avoid
owner: decision-coordinator after: set-review
automation: <level>
run-scoped:
outstanding - runs roll-call owner: decision-coordinator
every: <cadence>
from: take-in-trigger until: hand-over
accuracy - runs collect-and-report owner: analytics
every: <cadence>
from: cut-over until: run close
handoffs:
take-in-trigger -> read-scoring / read-exceptions / first-hand
[what-to-fix]: what the model has to fix, what each agent owes, and
the date it is owed. Three agents read one record
read-scoring -> gather-signals [scores-versus-outcomes]: the scores of
the period set beside what each lead went on to do, each score
naming the model version it was computed under
read-exceptions -> gather-signals [exception-decisions]: the leads no
rule could settle, what the person decided about each one, and the
field that was missing
first-hand -> gather-signals [first-hand-accounts]: what the people
who work the leads said, attributed to the person who said it
gather-signals -> clear-signals [candidate-signals]: each candidate
signal, the source it would be read from, and the terms that source
came with
gather-signals -> test-coverage [signals-to-test]: the same signals,
in the form they would be read in. Two branches read one record
clear-signals -> fit-candidates [cleared-signals]: the signals the
model may be fit on, and which contacts the consent record covers
for each one
test-coverage -> fit-candidates [coverage-findings]: how often each
signal is present, what the rest cost to fill, and which ones arrive
after the lead has been routed
fit-candidates -> back-test [candidate-models]: each candidate at a
version, with the data it was fit on and the assumptions it rests on
fit-candidates -> routing-effect [candidates-to-route]: the same
candidates. Two branches read one record
back-test -> choose-model [back-test-results]: what each candidate
would have scored the closed leads, and what those leads actually
did
routing-effect -> choose-model [routing-shift]: how many leads land in
each tier and each territory queue under each candidate, set beside
what the model in production does now
choose-model -> shadow-run [chosen-model]: the chosen candidate, why
it won over the others, and everyone who disagreed and on what
shadow-run -> issue-version [shadow-scores]: both scores for every
lead of the shadow period, and every lead the two models would have
routed to different owners
issue-version -> hand-over [versioned-model]: the model at its number,
and what changed from the version it replaces
hand-over -> cut-over [adoption-answers]: what each receiving agent
said it would change, and the date it said it by
cut-over -> watch-new-model [cut-over-record]: the date the new
version started scoring, and the last lead the version it replaced
scored
watch-new-model -> set-review [first-weeks-scores]: the first weeks of
scores set against what those leads did
set-review -> record-learnings [review-conditions]: the review date,
and the drift that brings the model back sooner
deviations:
clear-signals -> gather-signals [source-ruled-out]: legal rules out a
source, so the signals are gathered again without it
test-coverage -> gather-signals [signal-too-late]: a signal arrives
after the lead has been routed, so the signals are gathered again
back-test -> fit-candidates [no-candidate-beats]: no candidate beats
the model already in use, so the candidates are fit again on a
different set of signals
routing-effect -> fit-candidates [queue-flooded]: a candidate would
flood a territory queue, so the candidates are fit again or their
thresholds are changed
choose-model -> fit-candidates [room-cannot-agree]: the room cannot
agree on one candidate, so the candidates are fit again against what
the disagreement showed
shadow-run -> fit-candidates [shadow-disagrees]: the chosen candidate
behaves differently on live leads than it did on the closed ones, so
the candidates are fit again with the disputed leads attached
watch-new-model -> choose-model [scores-worse]: the new version scores
worse than the one it replaced, so the choice is made again and
whichever way it goes runs through the shadow, the version and the
handover
bindings:
roster: <who holds each role - agents claiming the abstract agents
above, and named people for the sales side, leadership and
legal>
systems: the CRM (read), enrichment sources (read),
analytics platforms (read),
the customer and revenue record (read),
the consent record (read), the model store (write),
the standards store (write), the decision record (write),
the review schedule (write),
the notification channel (write)
data: the model in production at its version, the territory map
at its version, <your lead record> over <what period>,
the terms attached to every source the model reads
policy:
every score records the model version and the inputs it was computed
from
a model in production is never edited in place. A change goes in as
a new numbered version, and the version it replaces stays readable
no candidate reaches the room without a back-test over the leads
that have already closed
a signal that cannot be read before the lead is routed is marked as
such, and <who> decides whether the model may score on it
a rate the model cannot source is reported as unsourced, and no
estimate is put in its place
no new signal is scored on until legal has signed that its source
may be used and the consent record has been applied to it
the model that goes into production is chosen by named people, and
the choice is never delegated to an agent
the people who work the leads sit in the room where the model is
chosen, and a model chosen without them is not issued
the run does not close until every agent that reads a score has said
what it changes, and an agent that changes nothing says why
measures:
cycle time: <target> from the trigger to the new version scoring
accuracy: <share> of the leads scored in the top tier that went on
to <what counts as a win>, read on the review date
coverage: <share> of the signals the model reads that are filled in
before a lead is routed
quality gate: no version reaches production without its back-test,
its routing effect and its shadow run
```
Take it somewhere
Use this process in CrewAI Flows
close
Paste this into an assistant that can read the web, such as Claude, ChatGPT or Cursor. It reads the specification and the current CrewAI documentation, then writes two files: the flow, and a note on what did not survive the translation. Read the note first. What a runtime cannot express is the part worth arguing about, and this process is a draft to argue with.
copy the prompt
406 lines · the document is inside it, so nothing else is needed
Convert the business process below into a runnable CrewAI Flow:
one Python file with a Pydantic state model, one Flow subclass, @start,
@listen, @router and a durable human feedback provider.
The document is a reference process written to the Agent Processes
specification. Read the specification before you start, because it defines
terms that look ordinary and are not:
https://agentcatalog.com/spec/agent-processes
Sections 6 (the phase graph), 6.5.1 (exception edges), 6.7.1 (deviations),
7 (automation) and 8 (handoffs) are the ones this conversion turns on.
Then read the current documentation for the primitives you will need, rather
than relying on what you remember of the API:
https://docs.crewai.com/en/concepts/flows
Flow, @start, @listen, @router, and_, or_, state, kickoff, plot
https://docs.crewai.com/en/learn/human-feedback-in-flows
@human_feedback, the provider protocol, from_pending and resume
https://docs.crewai.com/en/guides/flows/mastering-flow-state
@persist and what persistence actually promises, which is less than it sounds
WHAT THE DOCUMENT ASKS FOR
These hold wherever the process lands, and they matter more than style.
1. Each phase under `phases:` becomes one step, and keeps its name.
2. `after:` gives the edges. `after: a + b` is a join and waits for BOTH.
Reading it as "either" is the defect the specification calls out by name.
3. Every handoff carries a key in square brackets. Each key becomes one field
on the run's state, named exactly as the key with hyphens turned into
underscores, and the sentence beside it becomes that field's comment. The key
is the stable name; the sentence is prose that may be rewritten.
4. A phase MUST NOT begin before its inbound handoff exists. Where that is
checkable, check it in the step rather than assuming it.
5. `automation: never` is a gate a person signs. The run stops there and does
not continue until a person's decision comes back. Do not turn one into a
notification, a log line, or an automatic transition, whatever the queue
looks like.
6. Each line under `deviations:` is a backward or sideways edge, returning to
the phase named on the right. The key in brackets names it, and that name
belongs in the code.
7. A phase whose `after:` reads like "X or Y, whichever could not finish" is an
exception edge: it is entered when those phases FAIL, not when they succeed.
Do not wire it as an ordinary successor.
8. Anything in angle brackets is a blank the adopting organization fills in.
Leave each one as a named constant at the top of the file with a TODO. Do not
invent a value, a threshold or a date.
9. Record the document's `from:` line at the top of the file, so it says which
reference process and which version it was generated from.
10. Run-scoped lines under `run-scoped:` are work that runs alongside the whole
process rather than at one point in it, and a run may not close while one is
unfinished. Say in the code what you did about them, including if the answer
is that the runtime has nowhere to put them.
HOW THAT LOOKS IN CREWAI FLOWS
11. Build a Flow, not a Crew. A Crew is a team of roles with no graph, no join,
no persistence handle and no gate, and cannot express this document at all.
Each phase is one method on a single `Flow` subclass, keeping its name with
hyphens turned into underscores. A phase that genuinely needs role-based agents
builds its own Crew inside its own method body, and the Flow stays the graph.
12. Declare the state as a Pydantic model bound as the type parameter,
`class NegotiateTheAgreement(Flow[NegotiationState])`, and read and write
`self.state.field`. Never use the untyped dict form: the handoff keys are this
process's memory, and an untyped dict turns a misspelt key into a handoff that
is silently absent. Keep the auto-injected `id` field, which is what every
resume depends on.
13. `after: a` is `@listen(a)`. `after: a + b` is `@listen(and_(a, b))`. Never
write `or_` where the document writes `+`.
14. Check every inbound handoff at the top of the method and raise if one is
missing. `@listen` says when a method may run and says nothing about what is
in hand when it does.
15. The join is the trap, and it is proven rather than theoretical. `and_()`
empties its accumulator the moment it fires, so re-entering BOTH arms works
forever, and re-entering ONE arm after it has fired leaves it holding a single
trigger and waiting for the other for good. The cascade drains, CrewAI prints
that the flow completed, and `kickoff()` returns normally. So for any joining
phase that a deviation can send work back into, do not use `and_()` at all:
make the join a `@router` that both arms trigger, which reads the state fields
and emits its label only when every inbound handoff is present. A router is
re-evaluated against durable state every time and is never suppressed by the
once-fired set.
16. Do not write a phase as `@listen(or_(and_(a, b), "some_label"))`. There is
one accumulator per listener, shared across every branch of its condition and
wiped when any branch satisfies, so the label firing while the join is half
full erases the arm that had already arrived.
17. `automation: never` is `@human_feedback(message=..., provider=...)` stacked
under the method's `@listen`, with a provider whose `request_feedback` raises
`HumanFeedbackPending`. The run then persists, returns that object from
`kickoff()`, and a different process answers later with `from_pending(flow_id,
persistence)` and `resume(text)`. Do not take the default `ConsoleProvider`,
which calls `input()`: that gate exists only while somebody is watching a
terminal, and a run started by a scheduler either hangs or takes an empty
string.
18. Silence must not approve, and the platform's default is that it does. With
`emit=[...]` set, an empty resume collapses to `default_outcome`, or to the
first label when that is unset, with no model consulted and nobody named. Treat
an empty or unrecognised answer as a refusal in your own router. And write the
approver's name and the time onto the state yourself, because
`HumanFeedbackResult` carries the text, the outcome and a timestamp but has no
field for the person, which the specification requires.
19. Each line under `deviations:` is a `@router` named after the key, returning
a label named after the same key, with the target subscribing as
`@listen(or_(normal_trigger, "the_label"))`. Route rather than listen
directly, because a router is re-evaluated on every cycle while a top-level
`or_` listener is suppressed after it first fires.
20. An exception edge is a `try` and `except` around the failing phase's body,
recording the failure on the state and emitting an exception label from a
router. Wiring it as `@listen(or_(x, y))` fires when those phases SUCCEED, so
the clearing phase would run on every healthy run.
21. An unimplemented phase is a method with the document's own sentence as its
docstring and a bare `pass`, which is safe because listeners still fire on a
`None` return. An unimplemented `@router` is not safe: returning `None` emits
no label, every phase below it disappears from the run including the gates, and
the flow reports success. A stub router must return a hard-coded label with a
TODO beside it, or raise.
22. Turn persistence on with `@persist(SQLiteFlowPersistence(...))`, then treat
every phase that performs a real act as something that will run twice.
Persistence saves the state fields and nothing else, so a restart rehydrates
the data and runs the graph again from `@start`: a flow that had already sent a
written refusal sends a second one. Guard each acting phase with a state field
it checks and sets.
23. There are no timers, no deadlines and no cadences, and `run-scoped:` has no
counterpart at all. Put each `by:` value as a named constant, name the
run-scoped lines in the module docstring as unimplemented obligations, and say
in the fidelity note that no deadline in this document is enforced by anything.
These absences produce no diagnostic whatsoever, which is exactly why they have
to be written down.
Produce a second file alongside it, `FIDELITY.md`, and treat it as the more
important of the two. The code is for whoever builds this. The fidelity note
is for whoever has to decide whether this platform suits the process at all,
and that is usually a different person who will never read the code.
It has three parts.
**What came across.** Briefly: how many phases became steps, how many handoff
keys became state fields, which gates stop the run, which deviations became
edges. Counts and names, not reassurance.
**What did not, and what was done instead.** One entry per gap. For each one,
say what the document requires, what the platform can actually express, what
you did in its place, and what breaks if somebody later removes your
workaround. This last part matters most: a workaround nobody understands is a
workaround somebody deletes.
**What a person still has to decide.** The blanks are not a translation
failure, they are the point of a reference process, so list what has to be
filled in before this could run against anything real, and say which of those
choices the platform constrains.
Write it in plain English for somebody who has not read the specification, and
do not soften it. A translation of a reference process is a draft to argue
with, not a build artifact, and the honest account of what was lost is the most
useful thing you will produce.
Here is the process document.
```
PROCESS: lead scoring model id: <team>/lead-scoring-model v1
from: ref/mkt/lead-scoring-model v1
owner: <who> effective: <date>
trigger: the accuracy report on <your lead routing process> shows the
scores have drifted, or a new signal about a lead becomes
available, or the review date on the model in production
arrives
watch: record=<accuracy report>
system=<your lead routing process>
change=<the report shows the scores have drifted>
or watch: record=<candidate signal>
system=<the source the signal is read from>
change=<a new fact about a lead becomes available>
or watch: record=<model in production>
system=<your model store>
change=<the review date on the model arrives>
concurrency: one run at a time for one scoring model, because two runs
would put two models into production. Runs against
different models overlap, and lead routing keeps scoring
on the current version throughout
goal: one recalibrated model in production as a numbered version, with
the back-test, the routing effect and the shadow run behind it,
and every agent that reads a score told what changed
phases:
take-in-trigger - convenes briefing: what the model has to fix, who
decides it, what evidence each agent owes, and the
date the answer is due
owner: decision-coordinator after: trigger
automation: <level>
read-scoring - human: the scores of the last <period> set against
what those leads went on to do, split by tier
owner: analytics after: take-in-trigger
by: <days> automation: <level>
read-exceptions - system: the leads that could not be scored or
routed, and what a person decided about each one
owner: lead-scorer after: take-in-trigger
automation: <level>
first-hand - human: what the people who work the leads say
about the scores those leads arrived with
owner: researcher after: take-in-trigger
by: <days> automation: <level>
gather-signals - human: the facts that might predict a win, each
with the source it would be read from
owner: forecaster
after: read-scoring + read-exceptions + first-hand
by: <days> automation: <level>
clear-signals - convenes approval: legal signs that each new
source may be used to score a person, and the
consent record says which contacts it covers
owner: <your legal role> after: gather-signals
by: <days> automation: never
test-coverage - system: how often each signal is already filled
in, what the rest cost to fill, and which ones
arrive after the lead has been routed
owner: lead-scorer after: gather-signals
automation: <level>
fit-candidates - runs build-by-talent: two or three candidate
models, each carrying the data it was fit on and
the assumptions it rests on
owner: forecaster
after: clear-signals + test-coverage
by: <days> automation: <level>
back-test - system: each candidate run over the leads that
have already closed, to see what it would have
scored them and what those leads did
owner: forecaster after: fit-candidates
by: <days> automation: <level>
routing-effect - system: how many leads land in each tier and each
territory queue under each candidate
owner: lead-scorer after: fit-candidates
automation: <level>
choose-model - convenes bake-off: two or three candidates, one
chosen, with the reasons said out loud
owner: decision-coordinator
after: back-test + routing-effect
by: <days> automation: never
shadow-run - system: both models score every live lead for
<how long>, and only the version in production
routes anything
owner: lead-scorer after: choose-model
by: <days> automation: <level>
issue-version - human: the chosen model issued as a numbered
version, with the version it replaces left readable
owner: standards-keeper after: shadow-run
automation: <level>
hand-over - runs decide-and-announce: every agent that reads a
score hears the new version at once and says what
it changes because of it
owner: decision-coordinator after: issue-version
by: <days> automation: <level>
cut-over - system: the new version starts scoring every lead,
and the version it replaced stays readable
owner: lead-scorer after: hand-over
automation: <level>
watch-new-model - human: the first <how long> of scores set against
what those leads went on to do
owner: analytics after: cut-over
by: <weeks> automation: <level>
set-review - human: the date the model comes back up, and the
drift that brings it back sooner
owner: decision-coordinator
after: watch-new-model automation: <level>
record-learnings - convenes debrief: what to repeat, what to avoid
owner: decision-coordinator after: set-review
automation: <level>
run-scoped:
outstanding - runs roll-call owner: decision-coordinator
every: <cadence>
from: take-in-trigger until: hand-over
accuracy - runs collect-and-report owner: analytics
every: <cadence>
from: cut-over until: run close
handoffs:
take-in-trigger -> read-scoring / read-exceptions / first-hand
[what-to-fix]: what the model has to fix, what each agent owes, and
the date it is owed. Three agents read one record
read-scoring -> gather-signals [scores-versus-outcomes]: the scores of
the period set beside what each lead went on to do, each score
naming the model version it was computed under
read-exceptions -> gather-signals [exception-decisions]: the leads no
rule could settle, what the person decided about each one, and the
field that was missing
first-hand -> gather-signals [first-hand-accounts]: what the people
who work the leads said, attributed to the person who said it
gather-signals -> clear-signals [candidate-signals]: each candidate
signal, the source it would be read from, and the terms that source
came with
gather-signals -> test-coverage [signals-to-test]: the same signals,
in the form they would be read in. Two branches read one record
clear-signals -> fit-candidates [cleared-signals]: the signals the
model may be fit on, and which contacts the consent record covers
for each one
test-coverage -> fit-candidates [coverage-findings]: how often each
signal is present, what the rest cost to fill, and which ones arrive
after the lead has been routed
fit-candidates -> back-test [candidate-models]: each candidate at a
version, with the data it was fit on and the assumptions it rests on
fit-candidates -> routing-effect [candidates-to-route]: the same
candidates. Two branches read one record
back-test -> choose-model [back-test-results]: what each candidate
would have scored the closed leads, and what those leads actually
did
routing-effect -> choose-model [routing-shift]: how many leads land in
each tier and each territory queue under each candidate, set beside
what the model in production does now
choose-model -> shadow-run [chosen-model]: the chosen candidate, why
it won over the others, and everyone who disagreed and on what
shadow-run -> issue-version [shadow-scores]: both scores for every
lead of the shadow period, and every lead the two models would have
routed to different owners
issue-version -> hand-over [versioned-model]: the model at its number,
and what changed from the version it replaces
hand-over -> cut-over [adoption-answers]: what each receiving agent
said it would change, and the date it said it by
cut-over -> watch-new-model [cut-over-record]: the date the new
version started scoring, and the last lead the version it replaced
scored
watch-new-model -> set-review [first-weeks-scores]: the first weeks of
scores set against what those leads did
set-review -> record-learnings [review-conditions]: the review date,
and the drift that brings the model back sooner
deviations:
clear-signals -> gather-signals [source-ruled-out]: legal rules out a
source, so the signals are gathered again without it
test-coverage -> gather-signals [signal-too-late]: a signal arrives
after the lead has been routed, so the signals are gathered again
back-test -> fit-candidates [no-candidate-beats]: no candidate beats
the model already in use, so the candidates are fit again on a
different set of signals
routing-effect -> fit-candidates [queue-flooded]: a candidate would
flood a territory queue, so the candidates are fit again or their
thresholds are changed
choose-model -> fit-candidates [room-cannot-agree]: the room cannot
agree on one candidate, so the candidates are fit again against what
the disagreement showed
shadow-run -> fit-candidates [shadow-disagrees]: the chosen candidate
behaves differently on live leads than it did on the closed ones, so
the candidates are fit again with the disputed leads attached
watch-new-model -> choose-model [scores-worse]: the new version scores
worse than the one it replaced, so the choice is made again and
whichever way it goes runs through the shadow, the version and the
handover
bindings:
roster: <who holds each role - agents claiming the abstract agents
above, and named people for the sales side, leadership and
legal>
systems: the CRM (read), enrichment sources (read),
analytics platforms (read),
the customer and revenue record (read),
the consent record (read), the model store (write),
the standards store (write), the decision record (write),
the review schedule (write),
the notification channel (write)
data: the model in production at its version, the territory map
at its version, <your lead record> over <what period>,
the terms attached to every source the model reads
policy:
every score records the model version and the inputs it was computed
from
a model in production is never edited in place. A change goes in as
a new numbered version, and the version it replaces stays readable
no candidate reaches the room without a back-test over the leads
that have already closed
a signal that cannot be read before the lead is routed is marked as
such, and <who> decides whether the model may score on it
a rate the model cannot source is reported as unsourced, and no
estimate is put in its place
no new signal is scored on until legal has signed that its source
may be used and the consent record has been applied to it
the model that goes into production is chosen by named people, and
the choice is never delegated to an agent
the people who work the leads sit in the room where the model is
chosen, and a model chosen without them is not issued
the run does not close until every agent that reads a score has said
what it changes, and an agent that changes nothing says why
measures:
cycle time: <target> from the trigger to the new version scoring
accuracy: <share> of the leads scored in the top tier that went on
to <what counts as a win>, read on the review date
coverage: <share> of the signals the model reads that are filled in
before a lead is routed
quality gate: no version reaches production without its back-test,
its routing effect and its shadow run
```
Take it somewhere
Use this process in Google ADK
close
Paste this into an assistant that can read the web, such as Claude, ChatGPT or Cursor. It reads the specification and the current ADK documentation, then writes two files: the workflow, and a note on what did not survive the translation. Read the note first. What a runtime cannot express is the part worth arguing about, and this process is a draft to argue with.
copy the prompt
397 lines · the document is inside it, so nothing else is needed
Convert the business process below into a runnable Google ADK workflow:
one Python file with a Workflow, nodes, routed edges, a JoinNode,
RequestInput gates and a persisting session service.
The document is a reference process written to the Agent Processes
specification. Read the specification before you start, because it defines
terms that look ordinary and are not:
https://agentcatalog.com/spec/agent-processes
Sections 6 (the phase graph), 6.5.1 (exception edges), 6.7.1 (deviations),
7 (automation) and 8 (handoffs) are the ones this conversion turns on.
Then read the current documentation for the primitives you will need, rather
than relying on what you remember of the API:
https://adk.dev/graphs/routes/
nodes, tuple chains, Event(route=), JoinNode, back-edges
https://adk.dev/graphs/human-input/
RequestInput and the rerun_on_resume handoff
https://adk.dev/runtime/resume/
ResumabilityConfig, resuming by invocation id, at-least-once tools
https://adk.dev/graphs/data-handling/
Event.output against state, and the selector syntax in instructions
WHAT THE DOCUMENT ASKS FOR
These hold wherever the process lands, and they matter more than style.
1. Each phase under `phases:` becomes one step, and keeps its name.
2. `after:` gives the edges. `after: a + b` is a join and waits for BOTH.
Reading it as "either" is the defect the specification calls out by name.
3. Every handoff carries a key in square brackets. Each key becomes one field
on the run's state, named exactly as the key with hyphens turned into
underscores, and the sentence beside it becomes that field's comment. The key
is the stable name; the sentence is prose that may be rewritten.
4. A phase MUST NOT begin before its inbound handoff exists. Where that is
checkable, check it in the step rather than assuming it.
5. `automation: never` is a gate a person signs. The run stops there and does
not continue until a person's decision comes back. Do not turn one into a
notification, a log line, or an automatic transition, whatever the queue
looks like.
6. Each line under `deviations:` is a backward or sideways edge, returning to
the phase named on the right. The key in brackets names it, and that name
belongs in the code.
7. A phase whose `after:` reads like "X or Y, whichever could not finish" is an
exception edge: it is entered when those phases FAIL, not when they succeed.
Do not wire it as an ordinary successor.
8. Anything in angle brackets is a blank the adopting organization fills in.
Leave each one as a named constant at the top of the file with a TODO. Do not
invent a value, a threshold or a date.
9. Record the document's `from:` line at the top of the file, so it says which
reference process and which version it was generated from.
10. Run-scoped lines under `run-scoped:` are work that runs alongside the whole
process rather than at one point in it, and a run may not close while one is
unfinished. Say in the code what you did about them, including if the answer
is that the runtime has nowhere to put them.
HOW THAT LOOKS IN GOOGLE ADK
11. Build a `Workflow` from `google.adk.workflow`, and pin `google-adk>=2.0` in
a comment. Do not use `SequentialAgent`, `ParallelAgent` or `LoopAgent`: they
are deprecated in favour of the graph, and they carry their own defects around
state and control flow. The documentation moved to adk.dev, and anything you
remember about nesting agents rather than drawing a graph is out of date.
12. Each phase is one node keeping its name. Take the node kind from how the
phase resolves rather than from taste: a `human:` or `system:` phase is a plain
Python function node, and a `runs` or `convenes` phase is an `Agent`.
13. `after:` gives the edges, written as tuple chains in `edges=[...]`, and the
trigger is the `"START"` keyword. Take the order only from `after:` lines and
never from the order the phases are listed in.
14. `after: a + b` is a `JoinNode`, and it must be guarded, because this is the
worst trap of any runtime here. The join fires when every static predecessor is
marked COMPLETED, nothing ever un-completes a node, and stored outputs are
never cleared. So after a deviation re-runs one arm, the join fires the instant
that arm finishes and hands the next phase LAST PASS'S value for every arm that
did not re-run. It does not stall, it proceeds with stale data, and nothing
logs. Stamp each arm's output with a pass counter or a content hash, and have
the phase after the join compare the stamps and refuse to run when they
disagree.
15. Each line under `deviations:` is a routed back-edge: a router after the
phase on the left returning `Event(route=...)`, named after the key in
brackets, with one arm going back to the phase on the right and one going
forward. An unconditional cycle raises at construction, which is the one place
this model checks your work. Nothing budgets a routed cycle, so add your own
count and stop rather than looping forever.
16. Give every router an explicit `DEFAULT_ROUTE` arm, and route it to a phase
that stops and asks a person. A route value matching no key writes a log
warning, ends that branch, and lets the run finish reporting success with the
rest of the process never having happened.
17. `automation: never` is a `RequestInput` node of its own, never an `Agent`
asking a question. Decorate it `@node(rerun_on_resume=False)` and yield
`RequestInput(message=..., payload=..., response_schema=...)`, so the run
stops, persists, and delivers the person's answer to the node's successor as
its typed input. A resumed workflow runs its tools at least once, so any
irreversible act needs its own duplicate guard.
18. Make the gates durable or say plainly that they are not. Wrap the graph in
`App(..., resumability_config=ResumabilityConfig(is_resumable=True))` and pass
a persisting session service, never the in-memory one. Note in the file that
the command line and the web UI cannot resume a run, so whoever releases these
gates needs an operator surface that somebody has to write.
19. Every `Agent` in the graph gets `mode="single_turn"` and no `sub_agents`.
A non-empty `sub_agents` list silently adds a transfer tool, and a model that
uses it runs a different agent in this node's place while the graph's outgoing
edge fires on schedule regardless: the topology is honoured perfectly and the
work belongs to somebody else.
20. Model failure as a route, not as an exception. A node that raises does not
propagate: the failure is caught, recorded, and shuts the workflow down without
raising to the caller. So a phase that can fail catches its own failure and
returns `Event(route="could-not-finish")`, and the exception phase hangs off
that arm.
21. Keep every blank as a named module-level constant and never interpolate one
into an `instruction=` string. Angle brackets and curly braces are ADK's own
data selector syntax inside instructions, so a blank pasted verbatim stops
being a blank and becomes a selector.
22. `by:` and `not-before:` have no expression, and `@node(timeout=)` is not
one: it is an in-process wall clock that cancels the node and, because failures
are swallowed, ends the run silently rather than recording a missed deadline.
Nothing in `run-scoped:` has an expression either, and it must not be faked as
an ordinary node, because a node has to be reached and has to finish before
anything downstream starts, which is the opposite of what those lines mean.
Leave both out of the graph and name them in the fidelity note.
Produce a second file alongside it, `FIDELITY.md`, and treat it as the more
important of the two. The code is for whoever builds this. The fidelity note
is for whoever has to decide whether this platform suits the process at all,
and that is usually a different person who will never read the code.
It has three parts.
**What came across.** Briefly: how many phases became steps, how many handoff
keys became state fields, which gates stop the run, which deviations became
edges. Counts and names, not reassurance.
**What did not, and what was done instead.** One entry per gap. For each one,
say what the document requires, what the platform can actually express, what
you did in its place, and what breaks if somebody later removes your
workaround. This last part matters most: a workaround nobody understands is a
workaround somebody deletes.
**What a person still has to decide.** The blanks are not a translation
failure, they are the point of a reference process, so list what has to be
filled in before this could run against anything real, and say which of those
choices the platform constrains.
Write it in plain English for somebody who has not read the specification, and
do not soften it. A translation of a reference process is a draft to argue
with, not a build artifact, and the honest account of what was lost is the most
useful thing you will produce.
Here is the process document.
```
PROCESS: lead scoring model id: <team>/lead-scoring-model v1
from: ref/mkt/lead-scoring-model v1
owner: <who> effective: <date>
trigger: the accuracy report on <your lead routing process> shows the
scores have drifted, or a new signal about a lead becomes
available, or the review date on the model in production
arrives
watch: record=<accuracy report>
system=<your lead routing process>
change=<the report shows the scores have drifted>
or watch: record=<candidate signal>
system=<the source the signal is read from>
change=<a new fact about a lead becomes available>
or watch: record=<model in production>
system=<your model store>
change=<the review date on the model arrives>
concurrency: one run at a time for one scoring model, because two runs
would put two models into production. Runs against
different models overlap, and lead routing keeps scoring
on the current version throughout
goal: one recalibrated model in production as a numbered version, with
the back-test, the routing effect and the shadow run behind it,
and every agent that reads a score told what changed
phases:
take-in-trigger - convenes briefing: what the model has to fix, who
decides it, what evidence each agent owes, and the
date the answer is due
owner: decision-coordinator after: trigger
automation: <level>
read-scoring - human: the scores of the last <period> set against
what those leads went on to do, split by tier
owner: analytics after: take-in-trigger
by: <days> automation: <level>
read-exceptions - system: the leads that could not be scored or
routed, and what a person decided about each one
owner: lead-scorer after: take-in-trigger
automation: <level>
first-hand - human: what the people who work the leads say
about the scores those leads arrived with
owner: researcher after: take-in-trigger
by: <days> automation: <level>
gather-signals - human: the facts that might predict a win, each
with the source it would be read from
owner: forecaster
after: read-scoring + read-exceptions + first-hand
by: <days> automation: <level>
clear-signals - convenes approval: legal signs that each new
source may be used to score a person, and the
consent record says which contacts it covers
owner: <your legal role> after: gather-signals
by: <days> automation: never
test-coverage - system: how often each signal is already filled
in, what the rest cost to fill, and which ones
arrive after the lead has been routed
owner: lead-scorer after: gather-signals
automation: <level>
fit-candidates - runs build-by-talent: two or three candidate
models, each carrying the data it was fit on and
the assumptions it rests on
owner: forecaster
after: clear-signals + test-coverage
by: <days> automation: <level>
back-test - system: each candidate run over the leads that
have already closed, to see what it would have
scored them and what those leads did
owner: forecaster after: fit-candidates
by: <days> automation: <level>
routing-effect - system: how many leads land in each tier and each
territory queue under each candidate
owner: lead-scorer after: fit-candidates
automation: <level>
choose-model - convenes bake-off: two or three candidates, one
chosen, with the reasons said out loud
owner: decision-coordinator
after: back-test + routing-effect
by: <days> automation: never
shadow-run - system: both models score every live lead for
<how long>, and only the version in production
routes anything
owner: lead-scorer after: choose-model
by: <days> automation: <level>
issue-version - human: the chosen model issued as a numbered
version, with the version it replaces left readable
owner: standards-keeper after: shadow-run
automation: <level>
hand-over - runs decide-and-announce: every agent that reads a
score hears the new version at once and says what
it changes because of it
owner: decision-coordinator after: issue-version
by: <days> automation: <level>
cut-over - system: the new version starts scoring every lead,
and the version it replaced stays readable
owner: lead-scorer after: hand-over
automation: <level>
watch-new-model - human: the first <how long> of scores set against
what those leads went on to do
owner: analytics after: cut-over
by: <weeks> automation: <level>
set-review - human: the date the model comes back up, and the
drift that brings it back sooner
owner: decision-coordinator
after: watch-new-model automation: <level>
record-learnings - convenes debrief: what to repeat, what to avoid
owner: decision-coordinator after: set-review
automation: <level>
run-scoped:
outstanding - runs roll-call owner: decision-coordinator
every: <cadence>
from: take-in-trigger until: hand-over
accuracy - runs collect-and-report owner: analytics
every: <cadence>
from: cut-over until: run close
handoffs:
take-in-trigger -> read-scoring / read-exceptions / first-hand
[what-to-fix]: what the model has to fix, what each agent owes, and
the date it is owed. Three agents read one record
read-scoring -> gather-signals [scores-versus-outcomes]: the scores of
the period set beside what each lead went on to do, each score
naming the model version it was computed under
read-exceptions -> gather-signals [exception-decisions]: the leads no
rule could settle, what the person decided about each one, and the
field that was missing
first-hand -> gather-signals [first-hand-accounts]: what the people
who work the leads said, attributed to the person who said it
gather-signals -> clear-signals [candidate-signals]: each candidate
signal, the source it would be read from, and the terms that source
came with
gather-signals -> test-coverage [signals-to-test]: the same signals,
in the form they would be read in. Two branches read one record
clear-signals -> fit-candidates [cleared-signals]: the signals the
model may be fit on, and which contacts the consent record covers
for each one
test-coverage -> fit-candidates [coverage-findings]: how often each
signal is present, what the rest cost to fill, and which ones arrive
after the lead has been routed
fit-candidates -> back-test [candidate-models]: each candidate at a
version, with the data it was fit on and the assumptions it rests on
fit-candidates -> routing-effect [candidates-to-route]: the same
candidates. Two branches read one record
back-test -> choose-model [back-test-results]: what each candidate
would have scored the closed leads, and what those leads actually
did
routing-effect -> choose-model [routing-shift]: how many leads land in
each tier and each territory queue under each candidate, set beside
what the model in production does now
choose-model -> shadow-run [chosen-model]: the chosen candidate, why
it won over the others, and everyone who disagreed and on what
shadow-run -> issue-version [shadow-scores]: both scores for every
lead of the shadow period, and every lead the two models would have
routed to different owners
issue-version -> hand-over [versioned-model]: the model at its number,
and what changed from the version it replaces
hand-over -> cut-over [adoption-answers]: what each receiving agent
said it would change, and the date it said it by
cut-over -> watch-new-model [cut-over-record]: the date the new
version started scoring, and the last lead the version it replaced
scored
watch-new-model -> set-review [first-weeks-scores]: the first weeks of
scores set against what those leads did
set-review -> record-learnings [review-conditions]: the review date,
and the drift that brings the model back sooner
deviations:
clear-signals -> gather-signals [source-ruled-out]: legal rules out a
source, so the signals are gathered again without it
test-coverage -> gather-signals [signal-too-late]: a signal arrives
after the lead has been routed, so the signals are gathered again
back-test -> fit-candidates [no-candidate-beats]: no candidate beats
the model already in use, so the candidates are fit again on a
different set of signals
routing-effect -> fit-candidates [queue-flooded]: a candidate would
flood a territory queue, so the candidates are fit again or their
thresholds are changed
choose-model -> fit-candidates [room-cannot-agree]: the room cannot
agree on one candidate, so the candidates are fit again against what
the disagreement showed
shadow-run -> fit-candidates [shadow-disagrees]: the chosen candidate
behaves differently on live leads than it did on the closed ones, so
the candidates are fit again with the disputed leads attached
watch-new-model -> choose-model [scores-worse]: the new version scores
worse than the one it replaced, so the choice is made again and
whichever way it goes runs through the shadow, the version and the
handover
bindings:
roster: <who holds each role - agents claiming the abstract agents
above, and named people for the sales side, leadership and
legal>
systems: the CRM (read), enrichment sources (read),
analytics platforms (read),
the customer and revenue record (read),
the consent record (read), the model store (write),
the standards store (write), the decision record (write),
the review schedule (write),
the notification channel (write)
data: the model in production at its version, the territory map
at its version, <your lead record> over <what period>,
the terms attached to every source the model reads
policy:
every score records the model version and the inputs it was computed
from
a model in production is never edited in place. A change goes in as
a new numbered version, and the version it replaces stays readable
no candidate reaches the room without a back-test over the leads
that have already closed
a signal that cannot be read before the lead is routed is marked as
such, and <who> decides whether the model may score on it
a rate the model cannot source is reported as unsourced, and no
estimate is put in its place
no new signal is scored on until legal has signed that its source
may be used and the consent record has been applied to it
the model that goes into production is chosen by named people, and
the choice is never delegated to an agent
the people who work the leads sit in the room where the model is
chosen, and a model chosen without them is not issued
the run does not close until every agent that reads a score has said
what it changes, and an agent that changes nothing says why
measures:
cycle time: <target> from the trigger to the new version scoring
accuracy: <share> of the leads scored in the top tier that went on
to <what counts as a win>, read on the review date
coverage: <share> of the signals the model reads that are filled in
before a lead is routed
quality gate: no version reaches production without its back-test,
its routing effect and its shadow run
```
One run
A simulation of one run
The activities are on the left, whoever is doing the active one is on the right, and the record of the run builds up as it goes.
This run is built from the same rows as the diagram above: the left column is the activity list, the captions are the activity lines, the cast is the roster, and the labels on the wires are what the handoffs say actually passes.
Adoption
What you fill in
52 blanks to fill. Everything else is the process.
this process
from: ref/mkt/lead-scoring-model v1
Copy this line into your own document. It never claims this process is running anywhere; it records which draft yours started from, and it is what lets the catalog tell you when this one changes.
The header. Your own id, an owner, and the date it takes effect. One line records where it came from, and that line is what lets the catalog tell you when this reference process changes.
The roster. Which agent takes each activity, and which person takes each of the human ones. The process already names what it needs, so this is a lookup rather than a design exercise.
The numbers. Dates, budgets, cadences, and the targets in the measures block. Nothing here can be a reference value, because a target nobody chose is a target nobody meets.