The new rate is approved. Which quotes still use the old one?

A pricing committee approves a change for a defined segment. The internal rating service is updated. The project status turns green.

A broker portal still uses an earlier table. A delegated arrangement follows its own implementation timetable. An existing quote is renewed from a saved calculation. The business now has several versions of its pricing decision in circulation.

This is where a rate change needs an operational release plan. Approval establishes the intended decision. The channels need to apply it to the right business at the right time.

Here are six checks worth completing before treating the change as implemented.

1. Define which transactions are affected

Write down the applicable product, segment and jurisdiction, then distinguish new business, renewals and amendments where their treatment differs. Specify the effective-date logic and how already-issued quotes are handled under the applicable terms and rules.

Do not assume the date of data entry is the date that determines the rate. A quote may be prepared today for cover beginning later. An amendment can relate to an existing period.

The responsible pricing, underwriting and compliance specialists should approve these rules. An assistant can identify ambiguous cases; it should not invent the legal or contractual treatment.

2. List every place the decision is executed

Include the obvious rating engine and the less visible routes: broker interfaces, delegated systems, approved spreadsheets, manual referrals and batch renewals.

For each route, name the owner and the evidence required to confirm implementation. A central release does not prove that an external partner has changed its process.

An assistant can help compare documentation and identify references to old versions. Validate its inventory with the teams that actually issue business. The absence of a route from the documentation does not make it disappear from operations.

3. Test boundary cases

Use representative transactions around the effective date and across the affected channels. Include an existing quote, a new quote and a relevant amendment if those are in scope.

Keep the expected treatment and the actual result together. Testing only one newly created transaction can miss saved quotes or cached calculations.

Where two channels legitimately produce different results because of approved terms or authority, record the reason. Consistency means applying the intended rules, not forcing every number to match.

4. Trace version and authority

The issued result should be traceable to the approved rate or rule version and any authorised override. Preserve enough information to explain what was applied when the transaction was created.

An assistant may flag a result that differs from the expected version. It should separate a genuine implementation gap from an approved exception, an older valid quote or a data-quality problem.

That distinction matters commercially. Treating every difference as an error can trigger unnecessary corrections; overlooking a real release gap can leave unintended business flowing through a channel.

5. Reconcile after release

Sample actual transactions after implementation. Compare the route, effective dates, rule version and recorded exceptions. Ask partners for the agreed evidence rather than relying solely on a release acknowledgement.

If a channel is delayed, record its affected population and the temporary handling agreed by the authorised owners. Do not let “almost complete” conceal an unknown volume of business.

Prepare rollback or correction procedures before they are needed. A rollback must respect transactions already issued and the obligations they created. Reverting a table is not the same as reversing a contract.

6. Distinguish execution from pricing performance

Once the intended rules are applied, the next question is whether the pricing decision performs as expected. That requires appropriate exposure, claims development and portfolio analysis. Immediate consistency does not prove adequate pricing.

Keep the two reviews separate enough to diagnose a problem. A poor result may reflect an implementation defect, a changed risk mix or an assumption in the pricing strategy. They require different responses.

AI can assist with version comparison, exception grouping and reconciliation. Its value here is making the actual route of a pricing decision inspectable, including the channels the central system does not control directly.

Bring one recent rate change to a Strategy Session, with its channel list and implementation evidence. We can examine where the approved decision becomes difficult to verify.