Customer success as attestation reviews

After the close, the product is the customer’s accumulating attestation history — and customer success is the practice of reviewing it together. Nothing in this motion is a courtesy call; every touchpoint runs on evidence the customer already generates.

The adoption ladder: shadow → warn → enforce → strict

Enforcement is promoted deliberately, one mode at a time, so nothing breaks on day one. The modes come from the policy engine itself, and each promotion is a policy change the customer makes — which means the promotion is itself attestable. GROUNDED · Labs

Mode	What happens	When to promote
shadow	Policies evaluate and record; nothing is blocked.	Day one. Builds the baseline evidence with zero production risk.
warn	Violations surface to the customer’s review surface; traffic still flows.	Once shadow data shows the policy set matches real traffic — few false positives.
enforce	Non-compliant output is blocked, redacted, or refused before delivery.	When the customer trusts the policy set enough to let it act. This is where the risk-control story begins.
strict	The tightest posture — the declared baseline is the hard boundary.	Mature deployments with settled baselines and a working review practice.

The ladder is also the honest sales frame: “we start in shadow” defuses the break-production objection in one sentence, and the customer’s own shadow-mode evidence makes the case for each promotion. GROUNDED · Labs

Quarterly evidence packs and auditor bundles

The compliance value compounds on a rhythm:

• Quarterly evidence packs — the period’s attestation record summarized against the customer’s declared baseline and framework obligations; the artifact a governance buyer forwards to their board.
• Auditor bundles — machine-generated control-coverage exports an assessor can verify independently, which is the point: the audit-access design means the bundle’s authority doesn’t depend on Glacis’s interpretation. GROUNDED · Labs

Each pack is a renewal argument, an expansion argument (paths still outside the governed scope show up as the gap between packs), and audit-prep hours the customer didn’t spend.

Drift-event review choreography

A drift alert is the product working, and the review choreography should make that obvious rather than leaving the customer alone with a scary graph:

1. Alert — sustained deviation from the declared baseline surfaces as a trigger-grade event, independently of anyone’s manual reporting. GROUNDED · Brochure
2. Review the record — what was governed, what fired, what was blocked: the attestation history is the incident record; nobody reconstructs from memory.
3. Remediate — update the baseline (behavior changed legitimately), roll back the model or version, or tighten a control.
4. Close as evidence — the remediation itself lands in the attestation history. The event’s afterlife is a demonstrated response, not an unresolved anecdote.

Walk this loop with the customer the first few times. The lesson it teaches — detected, reviewed, remediated, attested — is precisely the behavior that will someday earn pricing credit, which is why CS choreography and the actuarial pathway are the same story at two altitudes.

Renewal = the customer’s own attestation history

“Credit for cleaner attestation records” GROUNDED · Brochure is the endgame of the whole CS motion: at renewal — of the Glacis relationship, and eventually of the customer’s own insurance program — the submission writes itself, because the evidence has been accumulating all year under an open standard. A year of low drift, healthy review completion, and promoted enforcement modes is a record nobody has to assert. The honest caveat stays attached: premium-relief outcomes for customers are not yet demonstrated (VERIFY) — the attestation history is the mechanism by which they would be earned, and that framing is the sale.

Next

→ 16 VERIFY hub: the checklist

Drill this page →17 bank questions stand behind what you just read. Check it while it’s warm.