How does Attio detect duplicate people?

Attio prompts you to merge person records it detects as duplicates using the email address attribute. Two records holding the same email are surfaced; records with different emails, or with no email at all, are not.

Why does Attio create duplicate people?

The most common causes are email sync creating a record from an inbound message before anyone checks whether the person already exists, imports run without mapping email as the unique attribute, form submissions using a personal address when the CRM holds a work address, and enrichment tools writing a new record rather than upserting onto the existing one.

How do you match duplicate people when the email has changed?

Combine signals instead of relying on one. An exact LinkedIn profile URL identifies a person across jobs. Failing that, use full name combined with a second independent signal — the same phone number, the same company, or a shared email local part such as jane.doe across two domains — and review those groups rather than merging them automatically.

Is it safe to merge people who share a name?

No. Common names collide constantly, and shared inboxes, family members and colleagues at the same company make name-only matching unreliable in both directions. Name similarity should produce candidates for human review, combined with at least one independent signal before a merge.

How to find duplicate people in Attio

Last checked September 1, 2026 · Deduptio is our product. Every claim about another tool comes from its public site or documentation and is linked in place — tell us if something is out of date.

The short answer

Match exact email first, then exact LinkedIn URL — both identify a person on their own. After that you are into combinations: full name plus phone, plus company, or plus a shared email local part. Review those. Never merge people on a name alone.

The structural problem with person records

A person's strongest identifier in a CRM is their email address, and their email address changes every time they change jobs. Everything awkward about deduplicating people follows from that one fact.

Attio surfaces duplicate people by exact email match and prompts you to merge. When both records share an address, that is the whole job. When the same human appears as jane@oldco.com and jane.doe@newco.com, there is no shared value anywhere in the two records except a name — and a name is not an identifier.

Where duplicate people come from

Email sync. Attio can create person records from inbound mail. Anyone who emails you and is not already in the workspace becomes a record, including people who already exist under a different address.
Imports without a mapped unique attribute. One CSV can double the object; see CSV import duplicates.
Forms and events. People sign up with a personal address, then talk to you from a work one.
Enrichment and outbound tools writing a create rather than an upsert — see Deduptio vs Clay.
Job moves. The single largest source in a workspace older than a year, and the one no exact rule catches.

Signals, ranked

Signal	Strength	Use it
Exact email address (normalized: lowercased, trimmed)	Very high	Alone. Merge in bulk after a sample check
Exact LinkedIn profile URL	Very high	Alone. Survives job changes, which email does not
Exact mobile phone number (normalized to E.164)	High	Alone in most workspaces; combine with a name if your data includes switchboard numbers
Full name + same company	Medium	Review. Good precision, and it is the common job-move case
Full name + same email local part on different domains	Medium	Review. `jane.doe@a.com` and `jane.doe@b.com` is usually one person
Fuzzy name similarity alone	Low	Candidates only, never an automatic merge
Same company + same job title	Very low	Not an identity signal. Two people can hold the same title

The false positives that matter

Shared inboxes. info@, sales@, hello@ and accounts@are not people. If several records share one, an email rule will merge a company's entire front desk into one person. Exclude role addresses from person matching.
Family members and namesakes. Shared surnames, shared addresses, sometimes shared phone numbers.
Common names. In a workspace of any size there are several of every common first-and-last-name combination.
Recycled phone numbers. Company mobiles reassigned between employees.
Assistants and shared calendars. One address representing two people.

Choosing which person record survives

Keep the record with the conversation. Email threads, calls and meetings are the hardest thing to reconstruct and the most valuable thing on a person record, so the record carrying the history usually wins even if the other one has fresher field values — copy those across instead.

Then think about the current email. If the person moved companies, the surviving record should end up holding the address they use today, with the old one retained if your data model allows multiple. And the company association should point at where they work now, not where they used to.

Preventing the next batch

Make every integration write with a unique attribute so it updates rather than creating.
Be deliberate about email sync auto-creation, especially during a migration — it is the most common cause of duplicates appearing mid-import.
Store the LinkedIn URL. It is the one identifier that survives a job change, and it turns the hardest duplicate case into an exact match.
Run a weekly scan so job-move duplicates surface while someone still remembers the person.

Scan your People object

Connect your Attio workspace, write one match rule, and run a scan. Scans are read-only — you see every group and the reason it matched before anything is merged.

Start a free Attio duplicate scan

Finding duplicate companies in Attio — the same exercise where the identifier is a domain.
Cross-object matching — using duplicate companies as evidence when matching people.
How to merge records in Attio · The full guide