Intercom Fin

Intercom Fin

Fin answers customer questions and hands unresolved conversations to support staff.

Oximy ResearchUpdated 19 September 2026

Overview

A+
Transparency
6 / 10
Trust
98 / 100
Chat assistantSan Francisco, USIntercom

Trains on your data

Free
No
Paid
Yes
Opt-out
Yes

Incidents

1
  • highCompromised version of intercom-client npm package and intercom-php
SOC 2ISO 27001GDPRHIPAACCPAISO 27018ISO 27701ISO 42001

Assessed SEP 21, 2026 · Checked by hand SEP 21, 2026

Fin asserts SOC 2, ISO, GDPR and HIPAA, excludes trial accounts from training, offers paid-customer training opt-out, and publishes a DPA, subprocessors and regional hosting.

Trust breakdown

A+

98 / 100

Overall trust score

Data safetyComplianceIncidentsLegal
Data safety100
Compliance100
Incidents90
Legal100

Data and privacy

Trains on your data

Free
No
Paid
Yes
Opt-out
Yes

On by default. Admins can turn it off, which is why it costs less than training with no way out.

Storage regions

USEUAustralia

Sub-processors

Amazon Web Services, Inc.Snowflake Inc.Mailgun Technologies, L.L.C.Message Systems, Inc. (d/b/a Sparkpost)Twilio, Inc.Cloudflare, Inc.OpenAI, L.L.C.PlanetScale Inc.Anthropic, PBCGoogle L.L.C.Microsoft CorporationEleven Labs, Inc.Ably Realtime Ltd.Cartesia AI, Inc.Intercom, Inc.Intercom R&D Unlimited CompanyIntercom Software Australia Pty LtdIntercom Software UK Limited

Data retention

Third-party AI providers are contractually restricted from using your data for training, with zero data retention policies in place once an output is generated.

Deletion: Upon termination or expiry of this Agreement, Intercom will (at Customer's election) delete or return to Customer all Customer Personal Data (including copies) in its possession or control.

Incidents

Powered by

Fin Apex 1.0Fin Apex FlashFin RetrievalFin RerankerFin Issue SummarizerFin Feedback ParserFin Language DetectorFin Escalation RouterOpenAIAnthropicGoogle
Proprietary model

Fin utilizes a suite of proprietary models (Fin CX Model Suite) developed by Intercom's AI team, including Fin Apex 1.0, Fin Apex Flash, Fin Retrieval, Fin Reranker, Fin Issue Summarizer, Fin Feedback Parser, Fin Language Detector, and Fin Escalation Router. These models are purpose-built and fine-tuned for customer service tasks. Additionally, Fin employs multi-model resilience across third-party LLMs from OpenAI, Anthropic, and Google.

Integrations and access

Integrations

API: YesIntercomFreshdeskSalesforce Service CloudHubSpot Service HubLinearStripeShopifyChatGPTClaudeGleanMakeMerge

Available on

Web
iOS
Android
macOS
Windows
Linux
Extension
CLI

Pricing

Base Plan

$49/mo

Includes 50 outcomes. Additional outcomes are $0.99 each. Resolutions, procedure handoffs, and disqualifications are $0.99 each. Qualifications are $9.99 each.

Enterprise

Custom

Minimum commitments apply. Contact sales for more details.

View pricing page

Prices as listed on SEP 21, 2026; check the vendor's page.

Moat and openness

7/ 10
Moat 7 out of 10

Moat

  • Proprietary Model
  • Proprietary Data
  • Network Effects
  • Switching Costmoderate
  • Unique UX
  • Distributionstrong

Fin demonstrates a strong moat primarily through its suite of proprietary, purpose-built AI models fine-tuned for customer experience, which likely leverage proprietary interaction data for continuous improvement. Its deep integration capabilities across various channels and existing tech stacks, combined with high resolution rates and an outcome-based pricing model, create significant value and moderate switching costs for businesses. While not having direct network effects, its comprehensive feature set and focus on enterprise solutions suggest a robust market position.

6/ 10
Openness 6 out of 10

Openness

  • Open model weights
  • Published research
  • Open source contributions
  • Transparency reports
  • Public safety evals

Intercom emphasizes trust and transparency in its AI products, particularly Fin. They have achieved several certifications, including AIUC-1 and ISO 42001, demonstrating their commitment to responsible AI governance, security, and data privacy. They provide customers with control over data used for AI training, offer detailed security and compliance documentation in their Trust Center, and have tools like 'Monitors' to help users understand and optimize Fin's performance. While they are transparent about their security practices and data handling, there is no indication of open-sourcing model weights or publishing general AI research beyond their product-focused blogs.

Timeline

  1. 2024-10Fin 2 launches on Anthropic’s Claude, with a stated 51% out-of-the-box resolution rate and 99.9% accuracy on retrieved content. It expands into multi-step reasoning, multi-language support across 45 languages, and introduces "Custom Answers".
  2. 2023-03Intercom launches Fin as an AI chatbot powered by OpenAI’s GPT-4, layered over its existing helpdesk.

Company

Fin

Founded
2011
HQ
San Francisco, California, U.S.
Site
fin.ai

Popularity

425.5K

monthly visits

medium tiermedium risk

Notable customers

AnthropicClayLightspeedRocket MoneyGamma

Fin is a leading customer agent product in a crowded field, known for its high resolution rates and outcome-based pricing model.

Value and ROI

What it should move: Fin resolves more questions at first contact and reduces resolution time.

Best fit

It fits teams that want an AI agent in their Intercom inbox.

The catch

Fin reads help content and customer conversations. Poor source material weakens answers, and the tool depends on Intercom.

  • Lightspeed Commerce: Uses Fin to resolve routine support questions before they reach an agent.[1]
  • Hospitable: Uses Fin to answer inbound questions from its support knowledge base.[2]

In our stacks for

References