Skip to content

Chat

Chat is Sonora’s interface over every connected system. You ask a question in plain English; Sonora figures out which sources to query, pulls the relevant rows and conversations, and answers with specifics — customer names, ticket IDs, call timestamps, exact quotes.

It is not a search box. It is the same investigative tool an analyst would build by hand, available to anyone on your team in seconds.

Three styles of question come up most often.

Investigation. “Why did Acme churn last quarter?” Sonora pulls renewal-period CRM history, support tickets opened in the 90 days before churn, recent meeting transcripts, and a usage trend. The answer cites the specific calls, tickets, and metrics behind each conclusion.

Filtering at scale. “Enterprise accounts with declining usage that mentioned a competitor on a call in the last 30 days.” Returns a list — names, ARR, the relevant call snippet, the usage trend — that you can refine in the same thread or export.

Operational lookups. “Open feature requests from customers above $100k ARR, grouped by request.” Returns the request, the customers asking for it, the date each one asked, and the original quote.

Follow-up questions stay in the thread, so “now filter to ones with renewals in Q2” or “which of these have an open Severity-1 ticket?” narrows the previous result rather than starting over.

Customer 360° is the full profile behind any account. Click a customer in a result and you land on their account overview and current health, every recent ticket and meeting (with sentiment), product usage trends, and the stakeholder graph. It’s the same view chat draws from — a starting point when you don’t have a specific question yet.

Threads have permanent links — drop one in Slack and a teammate sees the same answer with the same citations. Most lists export to CSV directly from the result. To monitor a question over time rather than ask it once, build a Canvas dashboard from it instead.

Anything that crosses your connected sources. Health, sentiment, churn risk, feature requests, competitive mentions, support backlog, expansion signals, champion engagement. Combine filters naturally: “enterprise accounts mentioning competitors with declining usage and an open ticket in the last week.”

If a question keeps coming up, save it as a custom prompt so the rest of the team can run the same query with one click.