You just signed up for a clariBI trial. There's a "Connect a Data Source" prompt staring at you. You've got 14 days, 50 AI credits, and no obligation to spend either. Here's how to make the first five minutes count.
Minute 1 — connect one source, not five
The instinct is to wire up every tool you own immediately. Resist it. Pick the single data source you already think about most — whichever tool's data you check most often. That's where the AI engine will pay off fastest because the questions you'll ask are already on your mind.
For most operators, that's one of:
- Stripe (live via MCP) — payments, customers, subscriptions.
- Google Analytics 4 (native OAuth) — web traffic and conversion paths.
- HubSpot (live via MCP) — CRM pipeline.
- Your production Postgres (direct DB connection) — whatever your app stores.
- A CSV export (file upload) — if your daily-driver tool isn't in the catalog yet.
Click Connect, authorize on the vendor screen, and clariBI will start preprocessing the source in the background. A "Data Insights" dashboard will materialize for that source within a minute or two.
Minute 2 — look at the auto-generated dashboard
While the preprocessing finishes, go to AI Dashboards in the sidebar. You'll see a dashboard titled "[Your Source] - Data Insights". This is auto-generated — clariBI's preprocessing engine inspected your data, identified the most likely metrics, and built a multi-widget dashboard without you configuring anything.
Spend 30 seconds clicking around. Notice the kinds of breakdowns it picked (time-series of the obvious numeric columns, category breakdowns of the obvious dimensions). This is the floor — your manual dashboards will be at least this good.
Minute 3 — ask one question in Conversational Analytics
Open Conversational Analytics. Type one question in plain English about the data you just connected. Examples:
- "What was my MRR last quarter compared to the quarter before?"
- "Which marketing channel had the lowest CAC last month?"
- "Show me top 10 customers by lifetime value."
The AI engine picks the right tool from your connected source, runs it, and shows you the answer with a chart. This is the moment the trial pays off — a question you wouldn't normally answer without an analyst is answered in 30 seconds.
Cost: 1–3 AI credits depending on complexity. You have 50.
Minute 4 — pin the chart you like
If the answer is useful, click "Pin to dashboard". The chart goes onto a dashboard you can return to or share. clariBI is keeping the result, not just the chart shape — the next time you open the dashboard, the data refreshes automatically.
Minute 5 — connect one more source
Now do it again. The second source can be more speculative — a system you'd like to be looking at more but haven't. The point of the first source was to validate the workflow. The point of the second is to surface answers across systems that no single tool could answer alone.
That's it. Five minutes, one data source connected, one AI answer pinned, one experiment in flight. Come back tomorrow with a more ambitious question and pin its answer too. By day five you'll have a dashboard of pinned answers — and you'll have spent maybe ten credits.
What not to do in the first week
- Don't connect every tool you own on day one. Each source you connect adds noise to the planner's tool selection. Add them as the questions come up.
- Don't try to replicate every report you currently run. Replicating known reports is fine on Lite. The trial's value is the questions you can't currently answer.
- Don't worry about credits. 50 credits is enough for ~30–40 ad-hoc questions. Most trial users finish the trial with credits to spare.
The full catalog is on the MCP integrations page. The auth model behind every connection is documented in our OAuth 2.1 explainer. Or just start the trial and skip the reading.