Skip to content
datatruth Get your Truth Score

Support · Intercom data quality for AIFin answers who it can see. Your data decides who that is.

Intercom sits at the front of conversational support, and with Fin it is increasingly the AI itself. That makes contacts, companies and conversations decisive. Leads, users and anonymous visitors split one person into several records, and custom attributes go stale. We compare Intercom exports with your CRM and billing to show what an agent would really know.

Get your Truth Score What to export

Starts with a 30-minute call. Commits you to nothing. Then exports only: no login, no connected app, nothing installed.

Illustrative example, not a client.

Source of truth

What Intercom gets right.

Every system is built to be right about something. Knowing what Intercom owns, and what it doesn’t, is where a cross-system check starts, and where the Truth Score™ (a score out of 100 for how far your data can be trusted) begins to take shape.

  • 01

    Conversations

    Messages, channels, assignment and resolution.

  • 02

    Product-side identity

    Users identified by external_id from your app.

  • 03

    In-app context

    Custom attributes and events your product sends.

What it isn’t the truth for

Account status. Whether a company is a paying customer, on which plan, is in billing and the CRM; Intercom attributes are a copy that ages.

Inside the system

Common Intercom data problems for AI.

What goes wrong inside Intercom itself, with the objects and fields where it shows. Each one is checkable from an export.

  1. 01

    Anonymous sessions never stitched

    Visitors who start as leads are never merged with the identified user, so one person’s history splits.

    roleanonymousemailexternal_id

  2. 02

    The same person as lead and user

    Inconsistent use of email and external_id leaves duplicate contacts across both roles.

    roleexternal_idemail

  3. 03

    Stale custom attributes

    Plan, seats or tier set once and never updated, so the agent reasons about the customer as they were.

    custom_attributeslast_seen_atupdated_at

  4. 04

    Weak company links

    Contacts not linked to the right company break account-level context.

    companiescompany_idcompany.name

  5. 05

    Conversations without useful tags

    Topics and tags applied inconsistently limit what can be learned from resolution history.

    conversation tagstopicsconversation_rating

Between systems

Intercom looks clean on its own. The gaps are next door.

The problems that sink AI sit between Intercom and the systems beside it: the same customer, person or transaction recorded differently in each. These are the gaps to test for each pairing.

Support

Intercom

Source of truth forConversationsProduct-side identityIn-app context

Intercom ↔ CRM

HubSpotSalesforce

  • Companies with no matching CRM account
  • Leads in Intercom never passed to sales

Intercom ↔ Billing and finance

NetSuiteXero

  • Plan attributes that differ from billing
  • Support for accounts that stopped paying

Intercom ↔ Product and warehouse

SnowflakeBigQuery

  • Users in the product with no external_id match
Illustrative: gaps typical between these systems, not client findings. Which ones exist in your data is what a Truth Score Read finds out.

What to export

What we need from you, in plain English.

Contacts and companies export to CSV from Intercom’s lists; conversation data comes from the reporting exports. We agree the scope first. No teammate seat, no app installed.

What to exportFor whoever runs it
Every contact, leads and usersContacts: id, role, email, external_id, created_at, last_seen_at, companies, key custom attributes
CompaniesCompanies: id, company_id, name, plan, custom attributes
Conversation metadata for the last yearConversations: id, created_at, state, tags, contact id, assignee, rating (text only if agreed)
  • Exports only, no access to your systems
  • NDA and data processing agreement first
  • The full list of sub-processors comes with our data processing agreement, before any data moves
  • Raw exports deleted after the engagement; we keep quality metrics, not your records

AI that depends on it

AI use cases that lean on Intercom.

If one of these is in a proposal or your plan, Intercom is one of the systems we compare. Each page sets out what the use case needs and where it breaks.

How a Truth Score Read tests Intercom

From your exports to a signed verdict.

Within 5 working days of complete exports. A fixed fee, agreed before we start, that never moves with the verdict.

  1. 01

    Export Intercom, and one system beside it

    The Intercom exports listed above, plus at least one neighbour, such as CRM. Under NDA, using the field list we send.

  2. 02

    Match across systems

    Our engine matches each contact and its related records across the exports by ID, name, email and domain, including the ones spelt differently in each system.

  3. 03

    Check by hand, then score

    Every red flag is checked by hand against sample records before sign-off. Then the Truth Score: each pillar shows what was measured and what wasn’t, and pillars with too little signal are marked, never guessed.

  4. 04

    Get the verdict in writing

    A signed memo: the red flags between Intercom and the other systems, a Leverability read on the AI you’re considering (which AI ideas you can use now, after fixes, or not at all), a verdict (sign, fix first or walk away) and a short fix list your own team can work from.

How the Truth Score Read works

What you get back

See what comes back: a real sample memo.

One fictional B2B software company, Example Co: 72,837 records in 7 exports from its CRM, finance and HR systems, run through the same engine and the same checks as a paid Truth Score Read. Each system passes its own checks. Matched across systems, only 13% of deals tie to an invoice.

Real engine output on generated test data, not a client.

Questions

Intercom and AI, answered plainly.

Anything else, ask on the 30-minute call. It commits you to nothing.

Book the call
Do you need to be added to our Intercom workspace?

No. Exports of contacts, companies and conversation metadata, run by your team, are enough. No seat for us.

We are switching on Fin. What should we check?

That contacts resolve to one person, that companies match paying accounts and that the attributes Fin reads are current. Those decide whether its answers fit the customer.

Why compare Intercom with billing?

Because plan and status attributes are copies. When they drift from billing, an agent can offer the wrong entitlements with confidence.

Do you read conversation content?

Only if a use case needs it, agreed in advance. The core checks run on structure and links, not text.

Find out what your Intercom data can carry. Before the AI does.

Your Intercom export and one other system in. The gaps between them in writing, within 5 working days of complete exports. Starts with a 30-minute call. Commits you to nothing.

Get your Truth Score See a real sample memo

Starts with a 30-minute call. Commits you to nothing.

Get your Truth Score