Reports
Historical reporting for a Twilio Flex contact center, in place of Flex Insights. Reports cover queues, agents, the IVR and bots, wrap-up outcomes, outbound calls, AI insights, coaching and cases, from one data model kept inside your own backend.
- Live data. Reports fold the same Event Streams deliveries as the wallboard, the moment they arrive. Today's numbers are seconds old, not an hour.
- Your definitions. Service level threshold, target and short-abandon threshold per queue, applied to history too. Custom report fields from any task attribute, without count or length limits.
- Before the task. Every Studio execution becomes an IVR or bot journey, joined to the contact, with no changes to your flows.
- One set of numbers. The wallboard, the Supervisor Console and the reports count with the same metric definitions, so the same metric over the same period is the same number everywhere.
- Open. Exports to Excel and CSV, scheduled emails, nightly Parquet and CSV files, a read-only SQL login for your own BI tool, Metabase, plain-language questions, and an MCP server for AI assistants.
- English and Hebrew throughout, including exports.
Who uses what
| For | Where | What |
|---|---|---|
| Supervisors and admins | Flex side navigation → Reports, or the reports console at https://<api-host>/reports/ | Standard reports, Explore, contact search with drill-down, Ask AI, settings |
| Agents | The same places | Their own numbers only: My performance, My time, and the contacts they took part in |
| Analysts | Metabase at https://<api-host>/bi/ | Ad-hoc questions and dashboards over the analytics schema |
| The customer's own BI | Read-only Postgres login; nightly files in the bucket | Power BI, Tableau or Looker on the fact tables; Parquet and CSV per day |
| AI assistants | https://<api-host>/analytics/mcp | MCP tools: list metrics, query, ask, search contacts |
How data flows
Each task's events are folded into segments (one row per queue visit and per agent leg) and contacts (one row per customer contact). Studio executions become journeys. Every minute, 15-minute interval tables are rolled up per queue and per agent. See Data model.
Compared with Flex Insights
| Flex Insights | Floor Walker Reports | |
|---|---|---|
| Freshness | Batch, about an hour | Seconds, same pipeline as the wallboard |
| IVR and bots | Not visible unless every flow creates a dummy task | Studio executions joined to the contact, plus milestones |
| Custom data | A fixed number of short custom attributes | Unlimited named fields from task attributes |
| Service level thresholds | Global | Per queue, set by supervisors, applied to history |
| Hebrew and right-to-left | No | Yes |
| Own BI or warehouse | Export API | Read-only SQL login, nightly Parquet and CSV |
| AI | Not included | AI analysis per interaction, questions in plain language, MCP |
| Matches the live views | No | Yes, one metric registry |
TaskRouter keeps 30 days of events, so the backfill can bring in the last 30 days. Older history stays in Flex Insights (or its export).
In this section
| Page | Covers |
|---|---|
| Using reports | The Reports view in Flex and the reports console: standard reports, Explore, Contacts |
| Metrics catalog | What each metric and dimension means |
| Service level settings | Per-queue threshold, target and short abandon |
| Saved reports and email schedules | Saving, sharing and emailing reports |
| Exports | Excel and CSV downloads, nightly open data files |
| IVR and Studio journeys | Journeys, exit types, caller input, milestones |
| Metabase and your own BI | The bi profile, the read-only login |
| Ask AI and MCP | Questions in plain words, the MCP server |
| Setup and backfill | Enabling reports, hold capture, custom fields, the 30-day import, retention |
| Data model | The analytics tables and how metrics are built on them |
| API | The /analytics endpoints |