# How to Get Your CRM Data Ready for AI | Dedupely

> AI assistants and agents read your CRM records, duplicates included. Find and merge duplicates before you roll out AI, and keep new ones out after.

- URL: https://dedupe.ly/knowledge-base/get-your-crm-ready-for-ai
- Section: knowledge-base
- Site index: https://dedupe.ly/llms.txt
- Any page: append `.md` to its URL for this plain-text form

---

## Getting Your CRM Ready for AI

By
[Gaby Salinas](https://www.linkedin.com/in/gbyslns)
Updated October 30, 2026
Merge duplicates before AI tools start reading, writing and acting on your CRM records.
AI assistants, agents and scoring tools work from the records in your CRM. When one customer is split across three records, AI sees three partial customers: it summarizes half the history, scores the wrong record, or contacts someone your team already talks to. In Dedupely, you get your CRM ready for AI by measuring your duplicates, merging the obvious ones, using Dedupely Intelligence for complex patterns, setting Merge Rules for the fields AI relies on, and turning on automation so new duplicates don't build up again.

### Before you start

-
**A connected and synced CRM.** See [Connect your CRM](https://dedupe.ly/knowledge-base/connect-your-crm).
- **A list of the objects your AI tools use**, usually Contacts, Companies and Deals (Opportunities in Salesforce and Close CRM).
-
**A yearly plan** for unlimited merging in Dedupely Intelligence. Monthly plans and the free trial merge with their available free merges. See [Dedupely Pricing](https://dedupe.ly/knowledge-base/pricing).

### Why do duplicates matter more once AI is involved?

| What AI does | What a duplicate causes |
| --- | --- |
| Summarizes a customer's history | It only sees the activity on one of the records. |
| Scores or prioritizes leads | Engagement is split, so a hot lead looks cold. |
| Writes or sends outreach | It contacts a customer as if they were new, or contacts them twice. |
| Answers questions about accounts | Counts and revenue are inflated or split. |
| Updates records on its own | It may update one record and not the other, so the duplicates drift further apart. |
A person often notices two records for the same customer. An AI tool reads each record on its own, across thousands of records at once.

### Step 1: How many duplicates do you have?

Start by measuring, so you know where to focus:
-
**Duplicate Report:** a weekly or monthly email with your duplicates across your connected CRMs. See [Notifications](https://dedupe.ly/knowledge-base/notifications).
- **Preset Duplicate Searches:** ready the first time you log in. Their match counts show which objects need the most work.
- **Dedupely Intelligence:** shows how many matches it found on Contacts, Companies and Leads, with nothing to set up.

### Step 2: Merge the obvious duplicates first

Exact duplicates are the quickest win and the safest merges. Use a search on at least three fields, like First Name, Last Name and Email on **Exact**. Test 1–2 Single Merges, check them in your CRM, then bulk merge 1–2 pages before you merge the rest. See [Your First Deduplication Project](https://dedupe.ly/knowledge-base/your-first-deduplication-project).

### Step 3: Use Dedupely Intelligence for the complex patterns

Dedupely Intelligence finds complex duplicates on Contacts, Companies and Leads without match settings. It's useful before an AI rollout because it catches the duplicates exact searches miss. Your CRM records are never sent to an AI model: Dedupely Intelligence runs as standard code inside Dedupely. See [Dedupely Intelligence](https://dedupe.ly/knowledge-base/dedupely-intelligence).

### Step 4: Decide which values AI should see

AI tools rely on a few fields more than others. Set [Merge Rules](https://dedupe.ly/knowledge-base/merge-rules)
for them before you merge in bulk:
| Field | Why it matters to AI | A common rule |
| --- | --- | --- |
| Lifecycle or lead status | Decides how a contact is treated | Keep the stage your process needs, for example the most advanced: one Win Rule per stage, in the order you want them to win |
| Owner | Decides who gets the task or alert | See [Merging Duplicates Without Breaking Ownership](https://dedupe.ly/knowledge-base/merging-duplicates-without-breaking-ownership) |
| Email opt-out | Decides who can be contacted | Keep the opt-out if any record has it: a Win Rule for the opted-out value, if the field isn't read-only in your CRM |
| Phone | Used for outreach | Keep the newest phone with a Win Rule on Create Date |
Win Rules run from top to bottom, so put the most important one first. Merge Rules only work on fields that aren't read-only in your CRM. Blank values never overwrite existing data, so a thin record won't erase a complete one.

### Step 5: Keep new duplicates out

Once AI is live, new duplicates keep arriving from forms, imports and integrations. To merge them soon after they reach Dedupely:
-
Turn on **Auto Merge** on your tested exact searches, with the **As soon as possible** schedule. See [Auto Merge](https://dedupe.ly/knowledge-base/auto-merge).
-
Turn on **Auto Match** on looser searches, so a person reviews them first. See [Auto Match](https://dedupe.ly/knowledge-base/auto-match).
-
Follow a routine. See [Your Deduplication Routine](https://dedupe.ly/knowledge-base/your-deduplication-routine).

### Example setup

A 40-person B2B team on HubSpot, on a yearly plan, is rolling out an AI assistant for its sales reps in a month.
- **Week 1:** checks the Duplicate Report and Dedupely Intelligence: about 6,000 contact matches and 900 company matches.
- **Week 1:** sets Merge Rules for Lifecycle Stage, Owner and email opt-out, then tests 2 Single Merges.
- **Week 2:** bulk merges its exact contact search (First Name, Last Name and Email on Exact), then companies on Company Name and Domain Root.
- **Week 3:** reviews Dedupely Intelligence results and merges the matches the team is sure of, and rejects the ones that aren't duplicates.
- **Week 4:** turns on Auto Merge (As soon as possible) on the exact searches and Auto Match on the rest, before the assistant goes live.

### Common mistakes

- **Rolling out AI first and deduping later.** AI tools start writing to records on day one, so duplicates multiply.
- **Rushing with loose searches.** A deadline is no reason to skip testing. Treat merges as permanent.
- **Only deduping contacts.** Duplicate companies split account history too.
- **Forgetting the fields AI reads.** Without Merge Rules, the wrong lifecycle stage or owner can survive.

### FAQs about AI and CRM duplicates

-
**Do duplicates affect AI tools in my CRM?** Yes. AI tools read each record separately, so a duplicate splits a customer's history, engagement and revenue.
-
**Is my CRM data sent to an AI model when I use Dedupely Intelligence?** No. It runs as standard code inside Dedupely's infrastructure.
-
**When should I start?** A few weeks before your AI rollout, so you have time to test your searches and review the less obvious matches.
-
**Will AI create new duplicates?** It can, if it creates records. Auto Merge on the "As soon as possible" schedule merges new duplicates soon after they reach Dedupely.

### Next up

[Dedupely Intelligence Find complex duplicates on Contacts, Companies and Leads, with nothing to set up.](https://dedupe.ly/knowledge-base/dedupely-intelligence)

### Still stuck?

Our team helps on every plan and during the free trial, by chat, email or Zoom.
-
Need help? [Talk to support](https://dedupe.ly/contact)
-
Questions about plans? [Book a Zoom with sales](https://meetings.dedupe.ly/meetings/gaby-salinas/talk-to-sales)
-
Reading this with a model? [Read this page as Markdown](https://dedupe.ly/knowledge-base/get-your-crm-ready-for-ai.md)
-
Feeding it the whole site? [Read llms.txt](https://dedupe.ly/llms.txt)
Behind the article

### Gaby Salinas

Marketing Assistant at Dedupely
Gaby Salinas came to CRM data from marketing and advertising, so she looks at duplicates from two sides: the match and merge logic behind them, and what they do to the people using the CRM, like the rep who calls a customer twice, the admin who gets blamed and the cleanup nobody wants to start.
As Marketing Assistant at Dedupely, she follows how RevOps teams, end users, consultants and technical teams actually work with their data: the records they stop trusting, the workarounds they build and the moment they finally decide to clean up. What she finds shapes most of what Dedupely publishes.
[Connect with Gaby on LinkedIn](https://www.linkedin.com/in/gbyslns)

### Get started for free

20 free trial merges and as much free support as you need to get your duplicates under control.
[Talk to sales](https://meetings.dedupe.ly/meetings/gaby-salinas/talk-to-sales)[Try for free](https://v2.app.dedupe.ly/signup)
[Or sign up with Google](https://v2.app.dedupe.ly/signup/google)
Running CRM cleanup for clients? [See the Partner Program](https://dedupe.ly/become-a-partner)
Skip to content
