Decision inputs
Facts that change the policy answer
For this request, user context, decision rules and product state is the input boundary and an automatic eligibility, routing or treatment decision is the output boundary. A useful check makes both explicit.
- 1Task and owner
- Product automation owner wants to make an automated product decision for a user. The request needs an accountable owner for an automatic eligibility, routing or treatment decision, even when the tool prepares most of the first draft.
- 2Information involved
- User context, decision rules and product state. Check uploads, history and connected systems before describing the request as low sensitivity.
- 3Tool and account
- An approved company account. Treat a new plug-in or connector as a change to the approved setup.
- 4Intended result
- The expected result is an automatic eligibility, routing or treatment decision. Its destination matters: private working material creates a different consequence from a sent, published or automated result.
- 5Consequence if it is wrong
- The decision can affect a person without explanation or meaningful appeal. A familiar task still needs escalation when this consequence becomes plausible.
- 6Human review
- product, risk and legal owners should inspect, change, reject or stop the result. The reviewer needs the source material and must be able to reject the output before it takes effect.
Possible policy routes
The task name alone cannot decide it.
A published workplace policy can return different answers for the same task. These are the practical branches worth encoding.
A routine policy route may be possible
The company can consider a standard route where the exact account is approved, only the minimum user personal information is used, an automatic eligibility, routing or treatment decision remains within the stated purpose, and product, risk and legal owners reviews it before use.
Approval may be required
A named reviewer should take over when the account or data handling is uncertain, the decision can affect a person without explanation or meaningful appeal, or an automatic eligibility, routing or treatment decision reaches people or systems beyond the requester’s authority.
The request may need to stop or change
Do not continue unchanged when restricted information would enter an unapproved service, the output would act before product, risk and legal owners can intervene, or define permitted inputs, explain outcomes and provide review for consequential decisions cannot be maintained. Consider less information, a controlled account or a non-AI process.
Request checklist
Questions to ask before using the tool
- 01
Which approved account will perform make an automated product decision for a user, and what external connections can it reach?
- 02
Could user context, decision rules and product state be reduced to a short de-identified extract?
- 03
At what point does an automatic eligibility, routing or treatment decision move beyond the requester’s private draft?
- 04
Does product, risk and legal owners have enough authority and time to stop the result?
- 05
When must the employee stop and run the policy check again?
Worked request
What the employee should submit
This example supplies decision facts without pasting the underlying material into the approval record.
- requester
- product automation owner
- task
- Use AI to make an automated product decision for a user.
- information
- user context, decision rules and product state
- tool
- An approved company account
- frequency
- Recurring work
- region
- Where the work and affected people are located
- purpose
- Automate action
- impact
- Automated user decision
- review
- Meaningful human review for consequential outcomes
- owner
- product, risk and legal owners
Useful safeguards
Controls that fit this request
- ✓
Define permitted inputs, explain outcomes and provide review for consequential decisions
- ✓
Start with a de-identified sample of user context, decision rules and product state before considering broader access.
- ✓
Write the boundary around an automatic eligibility, routing or treatment decision clearly so later users do not expand the approval by assumption.
- ✓
Link the completed check to the applicable policy version and append later reassessments separately.
Questions people ask
About this AI use
Is using AI to make an automated product decision for a user automatically allowed?
The task name cannot settle the answer. Apply the company’s published rules to user context, decision rules and product state, the exact account, an automatic eligibility, routing or treatment decision, its audience and the proposed review.
When is the request detailed enough to decide?
Describe an automatic eligibility, routing or treatment decision, identify user context, decision rules and product state, name the exact tool and account, explain who will receive or rely on the output, and state how product, risk and legal owners will review it.
What should remain after the decision?
Link the completed check to the applicable policy version and append later reassessments separately. A classification and controlled reference may be enough when copying user context, decision rules and product state would create unnecessary risk.