Tips for Better AI Responses

Give Alex or Crunchy one clear goal and the context needed to answer it. Check the selected organisation before asking questions about private business records.

Start with a useful question

Use a starter prompt, then add the location, dates and output you need. For example:

• Show me low-stock items at [location].
• Summarise purchase-order spend for [date range].
• Compare product gross profit percentage for [location] across these two weeks.
• Export those results as a CSV, including the dates and location.

Use BI Analysis for reporting questions. Include the measure you mean: stock value, purchase spend and gross profit percentage are different figures.

Ask for the evidence

Ask which records support an answer and whether any costs, dates or locations are missing. If a result seems wrong, name the specific item or record and ask Crunchy to check it.

An assistant can only use data and tools available to your organisation. A request does not unlock features outside your plan or fill gaps in your records.

Be explicit about changes

Say whether you want an explanation, a proposed change or an actual update. Review confirmations and onboarding drafts, then check the saved result in the relevant page.

For an import, inspect the mapping and validation results before approval. For an export, check the scope and totals before sharing it.

Keep each conversation focused

Use follow-up questions while working on the same goal. Choose New chat for a different task and Conversation history to return to earlier work.

Use the right assistant

Public Alex helps with discovery and setup preparation; it cannot read private Crunch records. Give it your Venue name and Postcode on Get Crunch for context.

Onboarding Alex helps prepare setup drafts for review. Dashboard Crunchy works with the selected organisation’s available tools and records. For exact steps, open the relevant Help article alongside the conversation.