How Rho cuts escalations 3× with continuous review of human and AI agents

Quality was in place.
Visibility was the missing piece.

Manual QA meant finding conversations, pulling them together and reviewing a small selection. That work covered only 3–5% of interactions, leaving most of the customer experience outside the review process.

For Stas Johnson-Chyzhykov, Rho’s Chief Operating Officer, the frustration was timing. The team was committed to quality, but often caught problems after they had already developed. Rho wanted to spot issues while there was still time to change the outcome.

From a small sample
to every interaction.

Rulebase expanded review to 100% of Rho’s phone, email and chat conversations, evaluating them against Rho’s operating standards. Managers could see when a conversation went off track and identify patterns across the wider operation.

The broader coverage gave the team a better basis for deciding what needed coaching and what needed an operational fix. A recurring issue could be investigated as a pattern, rather than appearing as an isolated miss in a small sample.

The same review covers AI agents. Rulebase reads bot conversations against those operating standards too, and reports resolution by intent rather than as a single average, so a missed intent, an unsupported answer or a failed tool call surfaces as clearly as a human miss.

Handoffs are judged, not just counted. A request that goes to a person is checked for whether it should have escalated and whether the specialist had the context to finish it, which separates an appropriate handoff from an avoidable one.

When a bot hands a request to a person, the review follows it through to resolution. That pairing is what starts the improvement loop: the human resolution shows the step the agent was missing, which becomes a knowledge, prompt or tool change, tested against realistic scenarios before it ships and measured on the same intent afterwards.

Less time assembling reviews.
More time going deeper.

For Jeff Pasquerella, Rho’s Head of Compliance, the difference was practical: gathering samples could previously take a week. Rulebase brought that work down to minutes, giving the team room to examine more conversations and investigate more thoroughly.

The team no longer had to be as selective about what it reviewed. That change helped surface issues that the old process would have missed.

What happens after
an issue is found?

Rho’s use of Rulebase extends beyond review. Its operations team runs agents that act on the signals: a customer showing signs of risk, another contact about the same unresolved issue, or a service deadline approaching.

Those agents bring in the responsible team and follow the work through. Weekly reviews also surface product feedback and service gaps, which are routed to the people who can address them.

In July 2026, Rho and Rulebase announced a multi-year partnership, extending the relationship from conversation review into automated customer-operations workflows.