Fin

What it does

Done-for-you service

PartlyYou run the software yourself, but a deployment team works alongside you on bespoke requirements.Our deployment team works directly with customers to optimize for their bespoke requirements, ensuring they are getting the best results possible.fin.ai ↗

How much it does unattended

Runs autonomously

YesBilling is defined by end-to-end resolution with no human involved; it escalates when confidence is low.Only pay when Fin resolves the issue end to end, or successfully executes a Procedure you've configured to end in a handoff to a human or a workflow.fin.ai/pricing ↗

Multi-agent

PartlyTwo agents exist: Fin faces customers, Operator runs operations and tunes Fin. The vendor does not describe agents collaborating on one conversation.Operator is an agent that runs your customer operations. It spots problems, finds opportunities to improve, and it acts on them – tuning Fin, keeping…fin.ai/operator ↗

Agent permissions

YesField-level data access restriction, OAuth-scoped integrations, and escalation instead of answering when it cannot answer safely.You choose what Fin can access. Select “Restricted data access” to limit Fin to specific fields, and rename internal values into customer-friendly…fin.ai/integrations ↗

Which models it runs on

Claude

YesAnthropic is one of the providers Fin routes across, though the vendor chooses when.Fin uses OpenAI, Anthropic, Google, and Intercom models, switching automatically to maintain performance.fin.ai/trust-reliability ↗

GPT

YesOpenAI is named in the same multi-model resilience list.Fin uses OpenAI, Anthropic, Google, and Intercom models, switching automatically to maintain performance.fin.ai/trust-reliability ↗

Open models

NoThe enumerated provider list is OpenAI, Anthropic, Google and Intercom's own models; no open-weight option is offered.Fin uses OpenAI, Anthropic, Google, and Intercom models, switching automatically to maintain performance.fin.ai/trust-reliability ↗

Model choice

NoThe vendor selects and switches models automatically; answers are generated by its own Apex model.Fin Apex 1.0 is the model that produces the final answer Fin gives to every customer.fin.ai/cx-models ↗

Where you use it

In a browser

YesHosted product configured self-serve in a workspace, deployed on Messenger, email, chat and more.Fin is the only Agent you can fully manage yourself, configuring tone, behavior, knowledge and more, all without needing any engineering resources.fin.ai ↗

Whose machine it runs on

Self-hosted

NoVendor-hosted SaaS; the only customer choice is which of three regions holds the data.Data hosted in the US, EU, or Australia based on residency needs.fin.ai/trust-reliability ↗

Open source

NoRuns on the vendor's proprietary, internally trained model suite; no source or weights offered.Fin performs best because it runs on our proprietary AI Modelsfin.ai ↗

What it costs to run

How it meters

YesPer outcome: $0.99 per resolution, Procedure handoff or disqualification; $9.99 per qualification.Resolutions, Procedure handoffs, and disqualifications are $0.99 each. Qualifications are $9.99 each. You're not charged when a conversation is…fin.ai/pricing ↗

Free tier

Partly14-day trial with unlimited outcomes and no credit card, but no free tier afterwards; the non-Intercom path is contact-sales.Free trials require no credit card to sign up. During the 14-day trial, you'll have unlimited access to Fin outcomes, with no limits on usage.fin.ai/pricing ↗

API access

PartlyFin Agent API and Data Connectors are publicly documented; the Fin API Platform (Apex, RAG, Retrieval, Reranker) is gated at $250k annual spend.The Fin API Platform is available to customers and prospects with annual spend of $250k or more.fin.ai/api-platform ↗

MCP server

YesMCP works both ways: Fin calls MCP-compliant tools, and external agents read Intercom conversation data over MCP.Yes. Fin supports custom integrations through REST API connectors, webhooks, and the Model Context Protocol (MCP). If your tool has an API, you can…fin.ai/integrations ↗

Not established: Bring your own key

Buying it for a team

A company can buy it

YesSold to organisations, including via sales with minimum commitments; 12,000+ business customers.Fin is used by over 12,000 of the world's most forward-looking brandsfin.ai ↗

Seat model

PartlyNo seat cost on third-party helpdesks, where teammates are unlimited; on Intercom, helpdesk seats are $29/month each on top of outcomes.Unlimited teammates – No additional seat costs.fin.ai/pricing ↗

Pooled budget

YesSpend is workspace-level outcome consumption, not per-seat, and is capped with configurable limits and alerts.Manage usage with configurable alerts, limits, and real-time reporting.fin.ai/pricing ↗

Admin controls

YesCustomer defines the agent's scope, boundaries and Procedures; data access can be restricted field by field.Fin operates within the scope you setfin.ai/trust-reliability ↗

Single sign-on

YesSSO via Okta, Azure AD and OneLogin, with SCIM provisioning; the specific protocol is not named.Fin integrates with Okta, Azure AD, and OneLogin for SSO; supports 2FA, SCIM, and IP restrictions to secure workspace access.fin.ai/trust-reliability ↗

Not established: Audit log

What happens to your code

Opt out of training

YesOpt out any time with deletion in 30 days; BAA, EU/AU-hosted and previously opted-out workspaces are excluded automatically.Fin uses anonymized customer data for fine-tuning. Opt out anytime—data is deleted within 30 days.fin.ai/trust-reliability ↗

Data residency

YesChoice of US, EU or Australia hosting.Data hosted in the US, EU, or Australia based on residency needs.fin.ai/trust-reliability ↗

Certifications

YesSOC 2 Type II, ISO 27001, ISO 27701, ISO 27018, ISO 42001, HIPAA, AIUC-1 and GDPR/CCPA, with documents self-serve.Fin holds SOC 2 Type II, ISO 27001, ISO 27701, ISO 27018, ISO 42001, HIPAA, and GDPR/CCPA compliance. All documentation is available to self-serve at…fin.ai/trust-reliability ↗

Not established: Getting out