AI Readiness Monitor
Are your Qlik apps governed well enough for an AI assistant to use them reliably? The AI Readiness Monitor solution answers that question app by app, with a score out of 100 and an action and remediation plan.
Measure, report, improve
The AI Readiness Monitor is a Qlik app built by HeyData. It reads your Qlik Cloud tenant’s APIs and scores each app with the same grid, at every assessment.
- Measure. Apps tagged “AI Readiness” are assessed by Qlik Automate, on a schedule or on demand. No manual input.
- Report. The Monitor shows the score, the maturity level, the action plan sorted by points to gain, the comparison between apps and the history.
- Improve. Our experts target the next level. Descriptions, glossary terms and data products are drafted with the help of AI, then validated by your teams. The next assessment confirms the fix.
A score based on evidence
- 17 checks per app, each with the detail of what was found.
- 100 points, spread across seven criteria.
- 83 points based on Qlik documentation or regulation (GDPR).
Seven criteria
- Master items · 20 points. Official measures and dimensions, which the assistant uses in preference to raw fields.
- Descriptions · 15 points. Business meaning spelled out: app, master items, fields.
- Linked glossary · 15 points. A shared vocabulary, linked to the measures. Only verified terms count.
- Data model · 15 points. No synthetic keys or isolated tables, technical fields hidden, no personal data exposed.
- Quality · 15 points. The Qlik Trust Score of the source datasets.
- Governed sources · 10 points. Sources taken from the catalogue and published in data products.
- Indexed for AI · 10 points. The app is available and indexed in Qlik Answers.
From score to maturity level
A level is reached when the app earns at least 75% of its points and the level below is reached. A high score is not enough: the foundations must be complete.
- Level 1 · Readable. The assistant explores the model on its own.
- Level 2 · Consistent. It calculates using your official definitions.
- Level 3 · Governed. It qualifies source, freshness and reliability.
Level 4, Agentic, is never awarded by the grid.
Rules that keep the score honest
- Never estimated. A check that no API can verify is flagged “not verifiable” and earns no points.
- No window dressing. Renamed fields, empty descriptions or unverified terms are not fixes.
- AI drafts, a human validates. Descriptions, glossary and data products carry business meaning: your teams validate them before they are written.
- A measurement, not an opinion. The quality of the assistant’s answers is measured separately, by accuracy tests replayed at every change.
Where do your apps stand?
We assess your first apps and walk you through the action plan.