CHATBOTS

Voice IVR knowledge-gap tracker for low-confidence answers

Captures every call where the ElevenLabs voice agent answered with low confidence or said it didn't know, clusters the failing questions in Postgres.

CategoryChatbots
Enginesim
Difficultyadvanced
Triggerwebhook
Steps5
Setup~25 min

How it runs

The automated pipeline, trigger to output.

  • TriggerElevenLabs posts call-ended webhook with confidence scoresElevenLabsElevenLabs
  • LogicFilter for low-confidence or unanswered turns
  • ActionUpsert question into Postgres knowledge-gap tablePostgreSQLPostgres
  • LogicCheck if gap crossed recurrence threshold
  • OutputFile Linear issue for recurring gapLinearLinear

What it does

Turns your voice agent's failures into a product backlog. Whenever the agent gives a low-confidence answer or admits it can't help, the question is logged and recurring gaps become tracked work items, so your FAQ coverage improves week over week.

When to use it

Use it when the voice bot is live but you suspect it's whiffing on questions you could easily document. Instead of guessing, you get a ranked list of the exact questions callers ask that the bot can't answer well.

How it works

  1. 1The ElevenLabs agent posts a call-ended webhook including per-answer confidence scores.
  2. 2The flow filters for any turn below the confidence threshold or flagged as unanswered.
  3. 3Matching questions are written to a Postgres gaps table and incremented if already seen.
  4. 4A logic step checks whether a gap has crossed the recurrence threshold this week.
  5. 5Newly recurring gaps create a Linear issue describing the question and frequency for the content team to address.

Set it up

What you configure once, before turning it on.

  1. 1
    Connect ElevenLabsText-to-speech, voice cloning.
  2. 2
    Connect HTTP webhookTrigger any URL on agent actions.
  3. 3
    Connect PostgresAny Postgres URL — query, write, migrate.
  4. 4
    Connect LinearIssues, projects, cycles, triage.
  5. 5
    Set each agent's modelWe leave models unset so you pick the tier — fast + cheap, or top-quality.
  6. 6
    Tune it to your dataEdit the prompts, filters, and field mappings so it matches how your team works.
  7. 7
    Test, then turn it onRun once against a sample, confirm the output, then enable the trigger.

Run this workflow in your colony.

14-day trial. No DevOps. No Sales call. Provisioned in under a minute.