What did we promise this customer that never shipped?
Every commitment made to an account, with delivery status: the exact thing the CSM gets asked on the QBR call.
- Best for
- QBR prep, escalation calls
- Primitives
- CommitmentsAccountsProvenance
- Token savings
- 59%
The prompt
You are preparing a commitments status for {{account_name}}. Using Sentra: 1. Pull every Commitment where the Account is {{account_name}} and direction is "from us, to them", across the full relationship. 2. For each, return: the commitment text, the date and Interaction it was made in, the named owner on our side, the promised timeline, and the current status (delivered, in-progress, slipped, dropped). 3. Group by current status. Output: - Section 1, "Delivered": short list, one line each. - Section 2, "In progress": item, owner, latest update Sentra has seen, expected date. - Section 3, "Slipped or dropped": item, original date, what changed (cite the Decision or Interaction where the change was made). - One paragraph: what to surface proactively before they ask.
See it work
Run it from
How it runs
- 01Copy the prompt into any agent connected to Sentra over MCP.
- 02The agent reads the company memory, scoped to what you are allowed to see.
- 03One grounded answer comes back, with every line traceable to its source.
Or automate it with Actions
The same prompt runs as a Sentra Action: on a schedule, or on a semantic trigger. Output lands in Slack or email, and every send waits for your approval.
Trigger idea: qbr prep, escalation calls
Harbor · promised vs. shipped
Delivered
commitment · closedSSO shipped Aug 2 against a Jul 15 promise. Confirmed in their admin console session.
Owed · CSV export
call · Jun 30 · 00:41:12"End of the month" said on the Jun 30 call. Never scheduled; asked about twice since.
Owed · API rate-limit raise
QBR · Jul 17Promised by Sofia in the July QBR. The ticket is still open with no owner.
Owed · Slack digest
threads · Jul 8, Aug 12Promised "next sprint" twice, five weeks apart.
The prompt above becomes this. Fictional data, real mechanics: every line cites the meeting, thread, or record it came from.
Without Sentra
- 01Pull every email, call transcript, and meeting note involving the account across the full relationship.
- 02Read each to identify promise-shaped language from our side. Re-read context to confirm intent.
- 03For each promise, identify the date, the speaker on our side, the promised timeline, and what was actually said.
- 04Cross-check delivery against shipped features, status updates, and follow-up threads.
- 05Manually group by status: delivered, in-progress, slipped, dropped.
- 06Find the Decision or Interaction where any slipped/dropped commitment was reversed — re-read internal threads.
- 07Compose the report.
~110,000 tokens
With Sentra
- 01Pull all `Commitments` where Account = customer and direction = from-us-to-them, across the full relationship.
- 02Each Commitment is typed with: text, originating `Interaction`, owning `Actor`, timeline, current status.
- 03Group by status. For slipped/dropped, the `Decision` that changed direction is already linked.
- 04Compose the status report.
~45,000 tokens