Salesforce agreed on June 15 to acquire Fin, the customer-service AI company formerly known as Intercom, for about $3.6 billion, giving Salesforce a more packaged support agent to put alongside its Agentforce platform.
The deal is expected to close in the fourth quarter of Salesforce’s fiscal 2027, subject to customary closing conditions and required regulatory clearances. Salesforce said the transaction is not expected to change its fiscal 2027 financial guidance or affect its capital-return program.
Fin’s main product is an AI agent built for customer support. It can handle queries across live chat, email, WhatsApp, SMS, phone, and Slack, and is powered by Apex, Fin’s proprietary model for support workflows. Salesforce said Fin’s agent has shown average end-to-end resolution of 76% of support volume and brings a customer base of more than 30,000 companies.
What Salesforce is buying
The acquisition is not just a talent grab or a generic AI feature add-on. Fin gives Salesforce a customer-support agent that is already sold as a focused product, with deployment patterns, customer-service integrations, and outcome claims that are easier for service teams to evaluate than a blank agent-building canvas.
That matters because Agentforce is broader. Salesforce has pitched it as the agent layer across Customer 360, where companies can build and govern agents tied to CRM records, workflows, data, and business processes. Salesforce said Agentforce reached $1.2 billion in annual recurring revenue in its fiscal first quarter of 2027, up 205% year over year.
Fin fits a narrower and more immediate buyer need: deflecting support tickets without forcing every service organization to design a custom AI-agent program from scratch. In practical terms, Salesforce is buying a faster on-ramp for companies that want support automation now, especially small and midsize businesses and commercial customers that need to launch quickly and connect to existing systems.
Why customer support is becoming the test case for AI agents
Customer service has become one of the clearest markets for enterprise AI agents because the work is measurable. A support agent either resolves the issue, escalates it, asks for the right information, follows policy, and leaves a usable record, or it does not. That makes the business case easier to test than more abstract promises about autonomous office work.
The workflow also gives AI vendors a defined operating area. Support agents can be connected to help-center articles, CRM records, order systems, account entitlements, escalation rules, and conversation histories. Those boundaries do not remove the risk of wrong answers or poor handoffs, but they make the product easier to govern than a general-purpose agent with open-ended access to company tools.
For Salesforce, the deal also helps answer a competitive question hanging over large software vendors: whether AI-native tools will sit on top of incumbent systems or be absorbed into them. Fin began as Intercom, built a customer-messaging business, then refocused its identity around AI support. Salesforce is now paying a large premium for that product-market evidence instead of relying only on internal Agentforce expansion.
The deal also points to a pricing shift
Salesforce has been buying pieces around the economics of AI software as well as the agents themselves. On June 8, it announced an agreement to acquire m3ter, a metering and rating platform for usage-based monetization, with plans to bring those capabilities into Agentforce Revenue Management. That earlier deal is smaller, but it is strategically connected to the Fin acquisition.
AI customer-service products are often judged by outcomes such as resolved conversations, successful deflections, handle-time reductions, and cost-to-serve changes. Those metrics do not fit neatly into old per-seat software pricing. If Salesforce wants Agentforce to become a durable business line, it needs not only agents that work, but also billing and measurement systems that let customers pay for usage, results, or a mix of both.
Fin’s own positioning leans into that shift. Its value depends less on putting another assistant inside a dashboard and more on whether it can close support interactions without creating expensive cleanup work for human teams. That is why Salesforce’s announcement emphasized measurable outcomes, fast deployment, and trusted agents rather than only model capability.
What customers should watch next
The deal still has to close, and integration will be the real test. Salesforce will need to decide how Fin remains available to existing customers, how quickly it is woven into Agentforce, what happens to Fin’s current product packaging, and how its Apex model fits with Salesforce’s broader model and data strategy.
Service leaders should watch for three practical details after closing: whether Fin can use Salesforce data without adding new setup complexity, how escalation and audit trails work when an AI agent touches customer records, and whether pricing is tied to seats, conversations, resolutions, or more granular usage. Those details will matter more to buyers than the language around the agentic enterprise.
The acquisition gives Salesforce a stronger answer in customer-service AI, but it also raises expectations. A $3.6 billion agent deal tells customers that support automation is no longer experimental software sitting at the edge of the stack. It is becoming part of the core CRM fight.
