Lorikeet launches Coach: an analytics agent to perk up AI customer support
Updated 09:00 EST . January 29, 2026
Lorikeet, an AI support platform for regulated industries, announced the general availability of Coach - a self-service analytics agent that evaluates support tickets and pinpoints where AI help-desk experiences are breaking down. It focuses on the visibility gaps that appear at scale: why ticket volumes spike, where sentiment drifts, and which topics, policies, or configurations trigger problems.
Customer satisfaction scores can hint at trouble, but they rarely explain the cause. "You've deployed AI support, but you're flying blind," said the company's CEO, Steve Hund. "CSAT tells you if customers are happy, but not whether the AI is doing what you told it to do." For teams that still live inside spreadsheets and sample reviews, that hits close to home.
If you need a quick refresher on CSAT, here's a straightforward explainer from Zendesk: What is CSAT?
What Coach does
Coach runs two capabilities in parallel: Thematic Analysis and Ticket Quality Score. One maps the "what" across your support stack; the other grades the "how" against your policies.
Thematic Analysis
Coach clusters conversations into meaningful topic buckets and then reports performance by topic. That view exposes what rollup metrics hide - the same AI might resolve 91% of one topic while struggling at 56% on another.
It also highlights shifts that matter to support leaders: sudden spikes, sentiment swings, and repeated misfires tied to specific flows or configurations. Instead of guessing, you get a short list of topics to fix first.
Ticket Quality Score
Every conversation is checked against a customizable scorecard and flagged on a traffic-light scale. Definitions align to your standards - reinforcing the app as the source of truth, avoiding certain refund scenarios, or preventing internal terminology from leaking to customers.
Traditional QA sampled a small slice of human tickets. That doesn't work when AI handles thousands of conversations a day. "The existing QA category was built for a world where humans answered every ticket and you could afford to sample," said the company's CEO. "But when AI is handling thousands of conversations a day, you need 100% coverage."
Coach evaluates both AI- and human-handled tickets, so you can hold a single bar for quality across channels. It's available to all customers and is offered as a standalone product.
Why this matters for support leaders
- See exactly which conversations fail - and why - instead of chasing vague scores.
- Prioritize fixes by topic impact, not gut feel.
- Enforce policy compliance with a scorecard that reflects your business rules.
- Track sentiment trends and volume spikes before they hit SLAs.
- Replace manual CSAT aggregation with actionable diagnostics.
How to put Coach to work this quarter
- Connect your ticketing data and map your core topics (billing, cancellations, onboarding, outages).
- Define a practical scorecard: source-of-truth checks, refund/discount guardrails, security and terminology rules.
- Set alert thresholds for topic-level drops or sentiment dips; review weekly trend reports.
- Run targeted fixes: adjust prompts, update policies, or tweak flows on the lowest-performing topics first.
- Close the loop with A/B tests and re-check performance after each change.
- Share a simple quality dashboard with Ops, Product, and Compliance to keep decisions aligned.
If your organization works under strict regulatory guidelines, consistent scoring across AI and human agents helps document adherence and reduce audit scramble. The shift is simple: move from sampling to full coverage, and from vibes to evidence.
Building team skills around AI support analytics? Explore role-based learning paths here: Courses by job. You can also browse the latest AI courses to level up support operations: Latest AI courses.
Your membership also unlocks:
