The AI tools getting the most attention right now, like Microsoft Copilot, Gemini in Google Workspace, custom chatbots and AI analytics, all have one thing in common: they work with your data. Your files, emails, spreadsheets and records. That’s what makes them useful. It also means that if your data is disorganized, outdated or overshared, AI will faithfully amplify those problems.
You don’t need a data warehouse or a data team to be ready. You need a few basics in place. Here’s what they are, and a quick assessment to see where you stand.
Key takeaways
- AI assistants can surface anything a user has permission to see. Tidy up sharing and permissions first.
- Outdated and duplicate files lead to confidently wrong answers.
- Know where your sensitive data lives before connecting AI to it.
- Pick one source of truth for important information like prices, policies and client lists.
Why data readiness matters more than the AI tool
Oversharing
AI finds what’s exposed
A folder shared with “everyone” years ago was effectively hidden. An AI assistant can find it and quote it in seconds, salary spreadsheets included.
Stale content
Old answers, stated confidently
If three versions of your price list exist, AI may pick the wrong one, and it won’t tell you it’s guessing.
Scattered data
Gaps in the picture
Information spread across personal drives, inboxes and USB sticks is invisible to business AI tools, and unprotected too.
How ready is your data?
Self-assessment
Data readiness score
For each statement, choose how true it is for your business.
The 30-day data clean-up plan
| Week | Focus | What to do |
|---|---|---|
| 1 | Find it | List where company data lives. Move important files off personal devices and drives into your business storage. |
| 2 | Lock it down | Review sharing: remove “anyone with the link” and company-wide access on sensitive folders. Remove ex-employee accounts. Turn on MFA everywhere. |
| 3 | Clean it up | Archive files untouched for 2+ years. Delete duplicates. Mark official versions of key documents and remove the rest. |
| 4 | Label and protect | Mark folders that hold sensitive data. Confirm backups work by test-restoring a file. Document who owns each key data source. |
Simple governance that doesn’t slow you down
- Name an owner for each key data source: client list, pricing, policies, financials.
- One home per document type. Decide where contracts, proposals and procedures live, and stick to it.
- Least access. People get access to what their role needs, reviewed when roles change.
- Retention rules. Decide how long to keep things, then actually archive or delete.
- Connect it to your AI policy. Your data classification (public, internal, restricted) should match your AI use rules.
Frequently asked questions
Can Copilot or Gemini see files people don’t have access to?
No. Business AI assistants work within each user’s existing permissions. The risk is that many users have access to more than they should, and AI makes that access far easier to use.
Do we need perfect data before using AI?
No. Standalone AI assistants used for drafting and research don’t touch your file stores at all. Data readiness matters most before connecting AI to company files, email and business systems. Start cleaning up in parallel.
Next step
See how ready your whole business is
Our free AI Readiness Assessment looks beyond data: strategy, skills, processes and culture.
