- What reaches a member Kept human
- Retention outreach, renewal correspondence, waivers and value statements are drafted by the system and sent by a person. An association is a relationship business and members can tell when something was automated at them. There is no volume at which this becomes automatic.
- Member data boundary
- What the workflow may see, what stays masked, how long it is kept and under which regime. Associations often hold employment, certification and incident information that members did not expect to be scored on.
- What a score may be used for
- A renewal risk score is a prompt for a conversation. It is not a reason to price differently, to withdraw a service, or to rank members to each other. That boundary gets written down before the first score exists.
- Certification integrity
- Who may amend a certification record, what evidence a change requires, and what the audit trail has to show. Where the association issues the credential, this is the one record with a regulatory consequence attached.
- Audit trail
- What was read, which rule produced which score, what was prepared, who reviewed it and what they changed. Built from the first workflow rather than added when the board asks how the scoring works.
- Escalation path
- What happens when a member is in dispute, in hardship, or in a category the rules did not anticipate. Every membership has all three, and a workflow that resolves them quietly is setting membership policy.
What counts as working. Accepted output against your own baseline is the
proof. The first build is usually tested against members who actually lapsed last year,
so you can see whether the scoring would have caught them before it is pointed at anybody
still in the building.
How we measure AI work sets out the method.