Use cases
Three worked scenarios that take you from a blank workspace to a real result. Each one is a click-by-click walkthrough you can follow in the hosted app, with the buttons to press and the screens you'll see.
New to AgentData? Do the Quickstart first (≈5 minutes) — it covers connecting a source, approving the model, and your first question. The scenarios below build on it.
| Scenario | Best for | What you'll build |
|---|---|---|
| AI lead enrichment → HubSpot | Startups, growth teams | A scheduled flow that enriches a company list with Apollo/Hunter and produces a HubSpot-ready list — no scripts. |
| Replace legacy ETL (SAP R/3 & friends) | Enterprise data teams | Read SAP read-only through an on-prem connector, transform with flows, and orchestrate them as a pipeline — instead of hand-maintained ETL jobs. |
| A unified semantic layer | Enterprise / regulated | One business model over many databases, queried in plain language from the app, REST and MCP — with row data that never leaves your org. |
Presenting AgentData to someone? Use the 10-minute demo script — a timed, click-by-click sequence (the wow, the privacy proof, the automation) with fallbacks.
Two ways to do everything
Every step is shown in the app and (where relevant) as the equivalent REST call, so you can demo by clicking or automate by scripting. See the API Reference.
Conventions used in these guides
- Buttons you click look like this: + Add source, ▶ Run now.
- Screens are shown inline as real app screenshots.
- Anything
monospacedis a value to type, a path, or a field name.