Your knowledge base can't hear your customers.
It only knows what your agents searched for. Knowledge Quest reads every conversation, finds the knowledge that's missing, stale or inconsistent, ranks it by impact and drafts the fix. Keep your platform. Make it first class.
Knowledge tools are only as good as the knowledge in them. Search, AI answers and agent assist all serve whatever is in the knowledge base. Knowledge Quest finds what's missing, out of date or contradictory from your real conversations, and drafts the fix for the knowledge tool you already run.Customer: "The storm took out half my fence. Do I still have to pay the excess?"
Agent: "Let me check with my team leader and call you back."
A search log records what agents typed. Your conversations record what customers need.
Legacy knowledge management finds gaps from failed and zero-result searches. That only catches a gap when an agent stops mid-call to search for it, in words the search happens not to match. Everything answered from memory, from a colleague or from a guess never shows up.
A handful of failed searches.
- "storm excess"
- "excess waiver"
- "ncb letter"
- "renewal increase why"
Agents who already "know" the answer don't search. Agents who are guessing don't search either. Neither shows up as a gap.
Every conversation.
- Storm claim excess waiver
- Premium increase at renewal
- Named driver change mid-term
- No-claims history letter
Every question a customer asked, matched against your knowledge base. Where nothing answers it, or two articles disagree, that's a gap.
Illustrative sample from a fictional insurer. Read more: Failed searches vs knowledge gaps.
Find the gap. Rank it. Fix it. Keep it fixed.
Four jobs, done continuously on the conversations you already have. Your team decides what gets published.
Find gaps in real conversations
Reads voice transcripts, chats, emails and tickets, works out what each customer wanted, and checks it against your knowledge base. Where nothing answers it, that's a gap.
Rank them by impact
Each gap is sized by how many conversations hit it and how hard they were: transfers, callbacks, unresolved contacts, negative sentiment. The highest-impact gaps come first.
Draft the fix with evidence
Knowledge Copilot drafts a new article or an update from the conversations themselves. Every draft links back to the transcripts it came from, and nothing publishes without your approval.
Keep knowledge accurate and consistent
Six quality checks run across your articles: brand alignment, readability, validity, duplicates, contradictions and complexity. Stale and conflicting content is flagged before customers find it.
The full method: Knowledge gap analysis, a practical guide.
We don't replace your contact centre platform or knowledge base. We make it first class.
Your agents know their desktop. Your team has years of articles. Knowledge Quest sits between the two: it reads conversations from one and sends better knowledge to the other. No rip-and-replace. You keep your tools.
Your contact centre platform
Conversations come in from Genesys Cloud, Amazon Connect, or your service desk tickets.
Knowledge Quest
Finds the gaps, ranks them by impact, drafts the fixes and checks the quality of what's there.
Your knowledge base
Approved articles and updates go back to the knowledge base your agents already search.
What you get: the gaps, their impact, and the proof.
A gap brief lists the knowledge your customers needed that your knowledge base couldn't give them, in order of impact, with the conversations behind every line. This one is a sample: the company is fictional and every number is invented to show the format.
| # | Gap | Conversations | KB searches | Est. cost / month | Recommended action |
|---|---|---|---|---|---|
| 1 | Excess waiver for storm claims | 1,140 | 9 | A$14.2k | Create articleDraft ready, cites 14 conversations |
| 2 | Why my premium went up at renewal | 1,310 | 4 | A$9.8k | Update existingTwo articles give different reasons |
| 3 | Changing a named driver mid-term | 690 | 31 | A$8.6k | Update existingArticle shows the old form |
| 4 | No-claims history letter for a new insurer | 420 | 0 | A$5.1k | Create articleDraft ready, cites 9 conversations |
| 5 | Roadside assistance on older vehicles | 305 | 12 | A$3.9k | Link existing articleAnswer exists; agents can't find it |
62% of these conversations were transferred, put on a long hold or ended with a callback promise. Sentiment fell in the second half of the call in 7 out of 10.
Customer: "Is the excess waived because it was the storm?" · Agent: "I'd have to check that for you."
Each figure links to the conversations it was counted from.
By impact: conversations affected, weighted up for transfers, callbacks, unresolved contacts and repeat contacts. The dollar estimate applies a sample cost per contact; your brief uses your own volumes and your own cost per contact.
Nobody can promise an AI is never wrong. We show our working.
Some knowledge vendors guarantee AI answers free from hallucinations. No one can honestly guarantee that. What we can do is make every claim checkable, so your team never has to take an answer on trust.
Every number checked
Every figure in a brief is checked against your interaction data and corrected before it's sent.
Every source cited
Gaps, drafts and answers link back to the conversations and articles they came from.
Deterministic scoring
Core scoring, alerts and operational metrics are rule-based and auditable. Language models draft and explain; they don't keep score.
Full provenance, people approve
Recommendations are traceable end to end, and every draft waits for a person to approve it.
Plugs into the knowledge base you already run.
Live knowledge sync with these knowledge bases today. Conversations come from Genesys Cloud and Amazon Connect; tickets from Freshservice and Zoho Desk.
Genesys Knowledge
Live knowledge sync
LivePro
Live knowledge sync
Freshservice
Live knowledge sync
Zoho Desk
Live knowledge sync
Use a different knowledge tool? Knowledge Quest works alongside other knowledge tools through import and export: bring articles in from PDF, Word, HTML or a web page, and take approved drafts back out.
Get a sample gap brief.
Tell us a little about your contact centre. We'll call to understand your queues and knowledge base, then show you what a gap brief from your own conversations would look like.
- Your gaps, ranked by impact. The questions customers asked that your knowledge base couldn't answer, and how many conversations each one affected.
- Evidence you can check. Every gap links to the conversations behind it.
- Nothing to replace. Read-only access to a sample of conversations. Your platform and knowledge base stay as they are.
Run a service desk on Freshservice or Zoho Desk? Start with the 30-day read-only ITSM pilot.
Prefer to talk first?
Call 02 9160 7760 or email info@customerscience.com.au.
Thanks, we have your request.
We will be in touch within one business day to set up a short call. If your email app opened as a backup, just hit send so the message reaches us.