San Francisco-based Lorikeet, an AI-powered platform specializing in regulated industries, has officially launched Coach, a self-service analytics assistant designed to evaluate support tickets and help organizations diagnose AI-related issues.
The company states that Coach addresses a critical challenge in AI support: as operations scale, visibility into problems becomes unclear. This includes understanding why metrics spike, why customer sentiment shifts, and which specific topics, policies, or configurations are causing issues.
While customer satisfaction surveys and scores (CSAT) can offer some clues, employees typically must manually read, aggregate, and diagnose the root causes. Lorikeet aims to go beyond this by providing enterprise users with a comprehensive, holistic view.
"You've deployed AI support, but you're still operating in the dark," said CEO Steve Hund. "CSAT can tell you if customers are happy, but it can't tell you whether the AI is performing the tasks you instructed it to do."
According to Hund, Coach's strength lies in being a diagnostic system. It precisely explains which conversations and topics failed and why, not just that they failed. This provides actionable direction for meaningful change.
Lorikeet explains that Coach operates with two core capabilities in parallel: Topic Analysis and Ticket Quality Scoring.
Topic Analysis uses AI clustering to categorize tickets into substantive themes, then breaks them down to generate performance metrics. The company says this reveals what aggregated metrics might hide; for instance, the same AI might achieve a 91% successful resolution rate on one topic but only 56% on another.
Ticket Quality Scoring examines each conversation and compares it against a scorecard, using quality checks to flag issues on a traffic-light scale. It utilizes a customizable set of definitions based on organizational standards—whether it's treating the application as a source of truth, avoiding certain refund scenarios, or ensuring the AI doesn't expose internal terminology.
This approach enables teams to move beyond traditional quality assurance tools, which typically sample only a small percentage of human-agent tickets or perform localized analysis. Instead, Coach allows for the examination of the majority of emerging support interactions.
"Existing QA categories were built for a world where a human answers every ticket and sampling is possible," Hund noted. "But when AI handles thousands of conversations daily, you need 100% coverage."
Interestingly, Lorikeet adds that Coach can process tickets handled by both AI and humans, allowing teams to maintain consistent quality standards across both domains.
The company confirms that Coach is now available to all customers and is offered as a standalone product.
