A data readiness checklist before you switch on AI
The questions to answer about your data before an Agentforce or Einstein pilot, and what happens if you skip them.
AI features are only as good as the records behind them. This checklist is what we work through during an AI readiness assessment.
Coverage
Are the fields the model needs actually populated, on most records, most of the time? Sparse data produces confident nonsense.
Permissions
Can the AI see only what the user is allowed to see? Get this wrong and you have built a data leak with a friendly interface.
Knowledge
Is your knowledge base current, or is it thirty articles last reviewed four years ago? Grounding on stale content is worse than not grounding at all.
Process fit
Is there a human step where a wrong answer gets caught? If not, add one before you go live, not after.
Want this looked at in your own org?
We will review your setup and tell you what we would change — no charge for the first conversation.
Book a free review