Customer support dashboards are good at telling teams something went wrong. They’re far less helpful at explaining why.
That gap is what Lorikeet is targeting with Coach, a newly launched AI agent designed to diagnose changes in customer experience metrics and push fixes into production. The company announced the product on Jan. 29, positioning it as an “AI co-worker” for support operations rather than another analytics layer.
The pitch is simple but ambitious: instead of combing through tickets after CSAT drops or response times spike, teams can ask Coach what changed — and let the system recommend, test, and deploy improvements.
Turning metrics into explanations
Support leaders typically rely on aggregate numbers like CSAT, first-response time, or resolution rate. When those numbers move, figuring out the root cause often means manually reviewing tickets, sampling conversations, or relying on inconsistent tags.
Coach takes a different approach. It evaluates every support interaction — whether handled by humans, AI agents, or a mix of both — and scores them against a customizable quality framework. Rather than surfacing charts, it responds to plain-language questions like “Why did CSAT drop last week?” with explanations tied directly to conversation data.
That conversational interface reflects a broader shift in enterprise software design. Instead of dashboards, vendors are betting that AI agents themselves become the primary interface — something you query, not configure.
Built for hybrid support teams
The timing is deliberate. Many companies now run hybrid support operations, where human agents and AI systems share workloads. Traditional quality assurance tools were built for sampling a small percentage of human conversations, not evaluating automated agents at scale.
Lorikeet argues that this creates blind spots — especially when AI introduces new failure modes, such as contradicting internal knowledge bases, exposing internal terminology, or drifting off-brand.
Coach applies the same quality standards across all conversations, allowing teams to compare human and AI performance directly. It also proposes fixes when it detects recurring issues, such as workflow changes or updated automation logic. For Lorikeet customers, those changes can be tested in simulation before being deployed.
Early signals from healthtech
One early user is HotDoc, a telehealth platform used by millions of patients. According to the company, compliance constraints and inconsistent ticket tagging previously limited its ability to extract meaningful support insights. By automating quality evaluation and thematic analysis, Coa