Build something that is not there yet
Build the headcount balance and its monthly roll-forward
- When HR, finance, and the board deck each report a different headcount for the same date.
- When nobody can reproduce last quarter's headcount because the HRIS only shows today's state.
Build an effective-dated worker snapshot and the point-in-time headcount and FTE by org unit, location, and employment type, with a monthly roll-forward that ties. Covers the headcount balance and its movements, not compensation or performance data.
Sample This Recipe has not been materialized in the Cookbook repository yet. Its trigger, description, prompt, agent guidance and acceptance conditions, and the explanation below, are prototype drafts. Its name, job, area and readiness come from the reconciled Cookbook seed snapshot. Readiness is a separate question from this one: it says whether the capability exists, not whether the writing has been reviewed.
Use this Recipe
Use the VibeData Recipe `people-headcount-rollforward` at https://getvibedata.ai/cookbook/people-headcount-rollforward Read the Recipe and execute it in the context of the current Intent.
Recipe id people-headcount-rollforward · Not yet materialized in the Cookbook repository, so the pointer addresses this page.
Verified by
What has to be observably true before this Recipe is finished.
- the roll-forward ties every month (opening plus hires minus leavers plus or minus transfers equals closing)
- an as-of date in the past reproduces the headcount figure reported for that date
- a worker with a transfer mid-period is counted in exactly one org unit as of any given date
- only the columns the recipe needs are landed from the HRIS source
Recipe promptThe task specification the agent reads. Reference only — it is not what you copy.
Deliver: Build the headcount balance and its monthly roll-forward. Build an effective-dated worker snapshot and the point-in-time headcount and FTE by org unit, location, and employment type, with a monthly roll-forward that ties. Covers the headcount balance and its movements, not compensation or performance data. Execute inside the current Intent. Its Domain, repository, platform, environment and attached sources are the context for this work — read them rather than asking for them. The work is done when: - the roll-forward ties every month (opening plus hires minus leavers plus or minus transfers equals closing) - an as-of date in the past reproduces the headcount figure reported for that date - a worker with a transfer mid-period is counted in exactly one org unit as of any given date - only the columns the recipe needs are landed from the HRIS source Report the evidence for each condition above with the result. A condition you cannot meet is something to say, not something to work around.
Agent guidanceHow the agent approaches the work, and what it will not do.
Profile the inputs the grain, joins and measures actually depend on before proposing a model. Put the design up for review — grain first — then build in an isolated copy with tests and documentation landing beside the model rather than after it.
Composes
- dbt model authoring and layering
- dbt in the project
- dlt in the project
- isolated-copy execution and gate verification
Asks first
Semantic decisions the Intent cannot supply. Never context Studio already holds.
- which Fabric target this work lands on, when the Domain carries both a Lakehouse and a Warehouse
- the grain the requester expects, where the request leaves it open to more than one reading
Guardrails
- Build in an isolated copy. Production is read, never written.
What you need
- Microsoft Fabric Lakehouse, Microsoft Fabric Warehouse, MotherDuck, or DuckDB.
- A dbt project you can build, and read access to the models it starts from.
- dbt and dlt in the project, or the intent to add it.
- Carries personal data; land only the fields the outcome needs.
How it goes
- State the outcome in one sentence, in the language the request arrived in.
- Let it profile the inputs the grain, the joins and the measures actually depend on.
- Review the design. Disagreeing about grain here costs a sentence; after the model exists it costs a rewrite.
- Let it build in an isolated copy, with the tests and the documentation landing beside the model rather than after it.
- Read the acceptance conditions against the run.
What you end up with
The deliverable, in your own repository, as transformation work a reviewer who knows the project reads as native to it. Alongside it, the evidence for every one of the acceptance conditions above — which is the part that is still there in three weeks when somebody asks.