Pre-built dashboards never match the question you have
Default analytics show what the vendor thought was important. Your actual questions — "how is the new agent doing on email this week vs last?" — aren't in the menu.
Type a question in plain English and get an answer back as a chart, table, and CSV — or build the query manually with a metric / group-by / filter picker. No SQL, no dashboards-by-committee, no export-to-Excel-hell.
Default analytics show what the vendor thought was important. Your actual questions — "how is the new agent doing on email this week vs last?" — aren't in the menu.
Custom queries get queued behind the quarterly board deck. By the time the answer arrives, the moment has passed.
Hand-rolled SQL access means injection risk, accidental cross-tenant leaks, and runaway queries. Most teams say no for good reason.
Every time you re-pull "top topics last 30 days" you redo the work. There's no notion of a saved, rerunnable report under your control.
"Show me avg response time by channel last week" → the AI translates to a structured ReportSpec, validates it against the allowlist, and runs.
Pick metric, group-by, time range. Multi-select channel/status, free-form topic and intent chips. The same engine the AI uses, with full control.
Date groupings render as line charts, categorical groupings as bar charts, ungrouped totals as a single big number. No chart-builder pixel-pushing.
Every metric, group-by, and filter is checked against a closed enum on the server. Queries are built with SQLAlchemy — no string-templated SQL, ever.
Export rendered rows as CSV in one click. Save any report by name to rerun with current data — no need to retype the question.
"Conversations by channel", "avg CSAT by agent", "daily volume trend", "bot vs human resolutions" — one-click templates so even week-1 users get value.
| Plain-English query → chart | Yes |
|
| No SQL, no injection risk | Yes |
|
| CSV export from rendered chart | Yes |
|
| Saved reports with one-click rerun | Yes |
|
| Included in base plan | Yes |
|
No. Type a plain-English question (e.g., 'show me avg response time by agent last week') and the AI converts it to a structured query. Or use the manual builder: pick a metric, group-by, time range, and filters.
No. Every query is automatically scoped to your organization. The query engine uses SQLAlchemy-built queries against a closed allowlist of metrics — no string-templated SQL, making cross-org data exposure impossible by design.
Yes. After building a report, click Save and name it. Saved reports appear in your list and rerun against current data on demand — no need to retype the question.
Volume (conversations, messages), response times (first response, resolution), CSAT scores, agent performance, QA scores, channel breakdowns, topic and intent distributions, SLA breach rates, and more. The available metric list grows with each platform release.
CSV export is available on every rendered report with one click. For BI tool integration (Looker, Power BI, Tableau), the AItocha API provides analytics endpoints you can connect to directly.
Saved reports are visible to all managers, supervisors, and admins in the same organisation. Per-user report access controls are on the roadmap.
Dashboard → Report Builder → click an example or type a question. The answer is a chart, a table, and a CSV — yours to save and rerun whenever.
Start Free Trial Blog