Healthcare & business ops workflows

Customer support FAQ gap-mining workflow

Your FAQ answers what you assumed users would ask, not what they're actually asking — and the gap shows up as support load.

Submit your version of this workflow Submit your version Subscribe
Audience
Healthcare & business ops workflows
Problem
Your FAQ answers what you assumed users would ask, not what they're actually asking — and the gap shows up as support load.
Outcome
Every week you publish 3 new FAQ answers grounded in real questions that didn't have an answer yet.
Tools used
  • Last week's support tickets or chat logs
  • An LLM in clustering-only mode
  • Your FAQ source
  • Plain text editor
Time required
1 hour per week
Difficulty
Easy

Steps

  1. Export last week's support questions with PII stripped (no names, no emails, no phone numbers, no IDs).
  2. Ask the LLM to cluster questions into themes and list the top 5 themes with example wording.
  3. Compare to existing FAQ. Flag themes with no current answer.
  4. Write the 3 highest-impact new FAQ entries yourself, not the LLM. Use exact user wording in the question.
  5. Publish, and add a 'last updated' date to each entry.

Example output

Theme 1: 'Can the bot pause replies during off-hours?' — 8 mentions, no FAQ entry. New FAQ: 'Yes, see the off-hours pause toggle in the host config.'

Human-review point. Human writes every FAQ answer. LLM only clusters and counts. No automated FAQ publishing.
Privacy notes. Never include customer names, emails, phone numbers, or specific case details. Use 'a member asked' phrasing.

Fork and remix ideas

  • Adapt for internal HR Q&A.
  • Adapt for an academic course FAQ.
  • Adapt for a community FAQ refreshed weekly.
related workflows

Keep going.

These pair well with this workflow.

Founder & operator workflows · Easy AI workshop intake template for teams

You're running an internal AI workshop and don't know what each attendee actually wants to build, so the session ends up generic.

 Founder & operator workflows · Medium WhatsApp community onboarding bot workflow

Every new community member asks the same 5 questions and you answer them at 11pm.

 No-code & automation workflows · Medium No-code lead qualification workflow

Inbound leads include great fits, casual browsers, and time-wasters — and you waste hours sorting them.

submit yours

Have a better version of this workflow?

Send the audience, the problem, the steps, and what you'd never automate. We review every submission with a human before publishing.

Submit a workflow Join the 4-week challenge