Skip to main content

Asking good questions

Exopen AI answers better when you're specific. Here are four patterns that lead to noticeably better answers — plus a few things the AI can't help with.

Be specific about the time period​

"Last quarter" and "last year" work, but "Q3 2024 vs Q3 2023" or "January–March 2025" gives the AI an exact period to compute on.

Worse: "How have sales been recently?"

Better: "Show revenue by month January to October 2024 vs the same period in 2023."

Name the customer, supplier, or project​

Use the actual name or number when you have it. The AI won't have to guess which customer or project you mean — and you avoid errors when several records have similar names.

Worse: "How are we doing with that big customer?"

Better: "Show receivables for Volvo Cars Sweden AB broken down by aging bucket."

For single figures, text is fine. But when comparing periods, customers, or products, or showing change over time, a chart makes the answer much easier to read.

Worse: "How has EBITDA developed?"

Better: "Show EBITDA by month for the last 24 months as a line chart."

Drill down with follow-up questions​

After the AI answers, you don't need to start over to go deeper. Ask a follow-up in the same conversation — the AI remembers what you just looked at.

Worse: Starting a new conversation and re-explaining the whole context.

Better: "Good. Now break that down by cost centre."

What Exopen AI can't help with​

Exopen AI can see your finance data. It cannot see:

  • HR or payroll data
  • Sales pipeline or CRM data
  • Real-time order status, inventory, or logistics
  • External market data or competitors
  • Anything that isn't in your finance database

If you ask about something outside finance data, the AI will say so — it won't make numbers up.

What now?​