The 30-Day Deflection Audit
A fixed-price, fixed-scope diagnostic that tells you exactly why your AI support agent under-delivers — and precisely what it would take to fix it. (“DeflectionDeflection the share of support conversations your AI resolves on its own, with no human agent involved.” is the number your vendor quotes; we measure the real one.)
We take 30 days of your real ticket logs, measure true containment against vendor-reported deflection, identify the top 10 intents your agent is failing, and hand you a costed remediation roadmap. If we can’t find at least a 10-point improvement path, you don’t pay.
- True containment vs. reported deflection, measured on your data
- Top 10 failing intents, ranked by ticket volume
- Root cause per failure — retrieval, routing, coverage or actions
- Costed, prioritised remediation roadmap
- Quick wins you can ship immediately
- Findings you keep, whether or not you build with us
Four weeks, four moves.
Instrument
Connect logs and knowledge base. Establish the real baseline your dashboard isn’t showing you.
Diagnose
Cluster failed conversations, isolate the top failing intents, and trace each to a root cause.
Prototype
Fix the highest-leverage failures in a sandbox so you can see the lift before committing.
Roadmap
A costed plan with expected containment gains, sequenced by effort and impact.
Your dashboard says 80%. Your customers know better.
Vendor “deflection” often counts any session that didn’t open a ticket — including the ones where the customer gave up. Measured containment across enterprises tells a different story, and the gap is where cost, CSAT and your credibility quietly leak.
The Deflection Audit replaces the vendor’s number with one you can trust, then shows you how to move it.
What exactly do I get?
A measurement of true containment vs. your vendor-reported deflection, a ranked list of the top 10 intents your agent is failing with root causes, and a costed remediation roadmap you own outright — whether or not you build with us.
What do you need from us?
Roughly 30 days of anonymised ticket / conversation logs, read access to your help centre and knowledge base, and 60 minutes with whoever owns the deployment. That’s enough to start.
What’s the guarantee?
If we can’t identify a path to at least a 10-percentage-point improvement in containment, you don’t pay. We’d rather tell you it’s already well-tuned than invent work.
Does the fee count toward a build?
Yes. The audit fee is credited in full against any remediation or build engagement that follows.
How long does it take?
Thirty days end to end, by design. You get interim findings inside the first two weeks.
Which platforms do you support?
Zendesk AI, Intercom Fin, Ada, Salesforce Agentforce, Sierra, Freshdesk, Kore.ai, Decagon, Gorgias and custom RAG stacks. If yours isn’t listed, ask — the diagnostic method is platform-agnostic.
Book your 30-day Deflection Audit.
Fixed price. Fixed scope. If we can’t find a 10-point path, you don’t pay.