Back to use cases
Product

Top customer pain themes this quarter

What customers complained about most, clustered and quoted: not your support ticket categories.

Best for
Quarterly roadmap reviews
Primitives
InteractionsThemesAccounts
Token savings
79%

The prompt

You are summarizing customer pain for the next planning cycle. Using Sentra: 1. Pull entities around customer-facing conversations and sentiment about us (sales calls, CS calls, support, community, etc). 2. Cluster the extracted pain points by theme. Do not use your own taxonomy — use the clustering the Semantic Agent has already learned for this organization. 3. For each theme, return: count of distinct accounts mentioning it, count of total mentions, three quoted moments with sources, and the most senior account that raised it. Output: - Top 8 themes, ranked by distinct-account count. - For each: one sentence on what customers actually mean, the three quoted moments, and the named PM or engineer who has already touched this area (per recent internal Decisions). - Closing paragraph: which two themes are new this quarter vs. recurring from last.

See it work

Run it from

ClaudeCursorChatGPTAny agent · MCP

How it runs

  1. 01Copy the prompt into any agent connected to Sentra over MCP.
  2. 02The agent reads the company memory, scoped to what you are allowed to see.
  3. 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: quarterly roadmap reviews

Sentraover MCPClustered from every customer conversation this quarter

Q3 pain, in their words

Export flexibility · 11 accounts

calls · 11 sources

"I end up screenshotting the dashboard into slides" is the modal quote. Raised by ops roles, not admins.

Time-to-value in onboarding · 7 accounts

onboarding calls

"Week one felt great, week three we forgot it was there." Concentrated in self-serve signups.

Permissions granularity · 5 accounts

security reviews

Enterprise only. Every mention pairs with a compliance review.

Roadmap check

Current quarter covers one of the three themes. Nobody has decided on exports.

The prompt above becomes this. Fictional data, real mechanics: every line cites the meeting, thread, or record it came from.

Without Sentra

  1. 01Pull every customer-facing interaction last quarter from CS, support, sales discovery, and community.
  2. 02Read each for pain-shaped moments — negative sentiment, feature gaps, blockers in their workflow.
  3. 03Tag each moment with a theme manually. Build the taxonomy on the fly.
  4. 04Aggregate counts per theme. Distinct accounts vs. total mentions.
  5. 05Pick the three most-quoted moments per theme — re-skim transcripts to verify verbatim.
  6. 06Cross-reference internal decisions to find the named PM or engineer who's touched each area.
  7. 07Compare against last quarter's themes to identify what's new vs. recurring.
  8. 08Compose the top-8 ranking with citations.

~280,000 tokens

With Sentra

  1. 01Pull `Interactions` where surface = customer-facing, last quarter, with `Decision`/`Rationale` polarity = concern.
  2. 02Cluster by `Theme` — Sentra's Semantic Agent has already learned the org's vocabulary.
  3. 03For each theme: count distinct `Accounts`, count total mentions, top three quoted moments with sources.
  4. 04Pull recent internal `Decisions` to identify the named PM or engineer who's touched each area.
  5. 05Diff against last quarter's themes — `Theme` trajectory is tracked.
  6. 06Compose the ranking.

~60,000 tokens

Book a DemoStart Free