How to run an annual policy attestation cycle with AI
An annual policy attestation cycle is the process of getting every employee — or a defined subset — to formally acknowledge they've read and understood specific company policies: a code of conduct, an acceptable-use policy, a data-handling policy, a conflicts-of-interest disclosure, and so on. Most operators run this once a year, sometimes more often if a policy changes. It's partly a legal requirement, partly an audit artifact, and partly the kind of hygiene that becomes very visible when something goes wrong.
The workflow looks like it should be easy to automate: you have a fixed list of policies, a roster of people, a deadline, and a binary completion status per person. That structure is exactly what makes operators reach for AI. The drafting work — policy summaries, reminder emails, manager escalations, completion dashboards — is high-volume and repetitive. A language model should be able to generate most of it. The instinct is right. The execution has real friction.
ChatGPT, Claude, and Gemini can contribute meaningfully to the drafting layer of this workflow. They'll write clear policy summaries, draft the initial attestation email and two or three follow-up reminders with escalating urgency, generate a completion-tracking template, and produce a final attestation report narrative. If you paste in the raw policy text, they'll summarize it accurately. What they can't do is send the emails, track who clicked, know who hasn't responded, or remember any of this when you run the cycle again next year.
How to do it with AI today
A practical walkthrough using ChatGPT, Claude, and other off-the-shelf LLMs — what they're good at, what you'll have to do by hand.
Where this gets hard
The walkthrough above works — until your numbers change, the LLM hallucinates, or you have to re-paste everything next month.
Tired of the friction?
Starch runs the whole workflow on live data — no copy-paste, no hallucinated numbers, no re-prompting next month.
The same workflow on Starch
Starch is an agentic operating system — it builds and runs the software your work depends on, connected to your live business data. For policy attestation, that means an agent builds the persistent app that manages the full cycle — drafts, sends, tracks, escalates, and reports — without you re-running prompts or reconciling spreadsheets.
Starch apps for this workflow
See this workflow by operator
The AI stack built for small in-house legal and compliance teams.
The AI stack built for small HR teams.
The AI stack built for small IT and ITOps teams.
The AI stack built for the founder's office.
The AI stack built for foundation and nonprofit ops teams.
The AI stack built for independent clinic owner-operators.
More AI walkthroughs in Compliance & Legal
SOC 2 audit evidence collection is the operational grind that sits between deciding to get certified and actually handing your auditor a complete evidence package.
Read guide →A Data Subject Access Request (DSAR) is a formal request from an individual — a customer, employee, or user — asking to see what personal data you hold about them, why you're processing it, and who you've shared it with.
Read guide →Responding to a subpoena or legal hold means identifying every relevant document, message, email, and record your business holds — then preserving it, logging it, and often producing it in a specific format under a hard deadline.
Read guide →Reviewing a vendor contract means more than skimming for price and term length.
Read guide →