Intercom No Longer Exists. Its AI Agent Took Over the Company Name.

On May 12, 2026, a 15-year-old technology company with 1,400 employees and 8,000 enterprise customers stopped existing, not through failure or acquisition but through a deliberate decision to replace its own identity with the name of its AI agent.

Not because it failed. Not because it was acquired. Intercom's CEO, Eoghan McCabe, decided that the AI agent the company had spent years building was so central to its future that keeping the old name sent the wrong signal about what the business actually was. So the company became Fin. That's the name of the AI.

McCabe's decision hands 1,400 employees a new parent company name and sends 8,000 enterprise customers a clear message: the AI agent isn't a feature of a customer messaging platform. The AI agent is the platform.

Why Intercom Bet Its Name on an AI Agent Called Fin

Intercom launched in 2011 as a customer messaging tool. It created a category that essentially didn't exist, built on the idea that companies should have real-time conversations with customers inside their products rather than routing them to a support email or call center. The product worked well enough that Intercom grew into one of the more successful enterprise SaaS companies of its generation, building a suite that covered sales, marketing, and support messaging across a single platform.

Then AI arrived at scale, and the company faced a problem common to established brands in a fast-moving category: newer competitors were taking market share with less capability and less track record, partly because they had no history to explain away.

McCabe put it plainly in the company's official announcement: "The relative success of the newcomers in our category is a result of the fact that they have no baggage." Companies starting fresh in 2024 or 2025 could position themselves as AI-native from day one. Intercom had 15 years of associations tied to live chat, sales automation, and product tours, features that look, to a prospect evaluating AI customer agents, like legacy constraints rather than core capabilities.

The name change is an attempt to cut that association at the source. Instead of trying to shift what "Intercom" means, McCabe let the Intercom product name persist for the help desk platform while moving the parent company identity to Fin entirely. All 1,400 employees now work for a company called Fin. McCabe acknowledged in the announcement that the company should have done this sooner.

The numbers behind the bet are substantial. Fin resolves more than 2 million customer issues every week. Those customers include Anthropic, DoorDash, and Mercury. Anthropic deployed Fin and achieved a 50.8% resolution rate within roughly six weeks, with Fin participating in 96% of all support conversations and saving the team more than 1,700 staff hours. Sharesies, a New Zealand investment platform, hit a 70% resolution rate within 12 weeks. Fundrise saw Fin resolve more than half of all support cases within three months.

The average resolution rate across Fin's full customer base sits at 67%, a figure the company says has increased every month as the model handles more complex queries. Fin 2, the current version, claims a 99.9% accuracy rate and handles more than half of all customer queries without any human involvement. The pricing model charges $0.99 per resolved conversation. For a team handling 10,000 support tickets per month, that's roughly $6,630 for the 6,700 Fin resolves without a human, compared against the full cost of staffing and managing agents to handle the same volume manually. In many deployments, the economics work substantially in Fin's favor. That math is why McCabe believes Fin is about to become the largest revenue driver the company has.

Fin Operator and the AI That Manages the AI

There's a part of customer service AI that doesn't appear in resolution rate dashboards, and it's the part that Fin's newest launch is designed to address.

An AI agent like Fin doesn't work out of the box and stay working without ongoing attention. It draws answers from a knowledge base that needs constant maintenance as products change, new features ship, policies update, and edge cases accumulate. When a company releases a significant product update, every piece of documentation Fin uses to answer questions about that product needs to be reviewed and potentially updated. When Fin starts giving wrong answers in a specific category, someone needs to diagnose whether the problem is the model itself, a knowledge gap, or a change in the product that nobody reflected in the docs.

Performance monitoring adds more work. When resolution rates drop over a two-week period, someone needs to determine whether the query mix changed, whether a product launch created a flood of questions the agent wasn't ready for, or whether documentation for a key area has gone stale. These are not automated answers. They require a human analyst to investigate, interpret, and act. At the scale of 8,000 customers and 2 million conversations per week, this operational load is substantial. The people doing it are support operations specialists whose job is essentially: keep the AI working well.

Fin Operator, announced May 15, 2026, is an AI agent whose entire job is managing this process. It entered early access for Pro-tier customers on the announcement date, with general availability planned for summer 2026. Operator fills three distinct roles that previously required significant human support operations time.

The first is data analyst. Support ops teams frequently need to know how Fin performed during a given period. Operator handles questions like "how did my team perform last week?" by generating on-demand charts, trend analyses, and drill-down reports from data already stored in the platform. Instead of a human spending an afternoon in dashboards, Operator produces the analysis on request.

The second is knowledge manager. When a product update ships, Operator can ingest the update announcement, search the entire content library for affected articles, identify documentation gaps, draft new content, and suggest edits to existing articles in a diff-style review interface. The work of identifying, researching, and drafting moves to the AI; the work of reviewing and approving stays with the human.

The third is agent builder, helping configure and tune Fin's behavior for specific scenario types where the agent's responses don't match company policy.

The Operator launch carries an implication worth examining carefully. Fin's core argument for years has been that AI can replace the humans answering customer questions. Operator is an acknowledgment that replacing human agents doesn't eliminate operational complexity; it relocates it. Instead of managing the humans who answer questions, support operations teams now manage the AI that answers questions. That management job is complex enough that Fin decided to build a second AI to handle it.

This maps to a broader pattern in enterprise AI adoption. Recent data showed that 74% of companies pulled their AI agents after initial deployment, with reliability concerns and the operational burden of maintaining agent quality among the most common reasons. The problem isn't usually the AI itself; it's the ongoing work of keeping the AI correctly calibrated against a changing business environment. Operator is Fin's argument that this ongoing work can be substantially automated.

Competitive Positioning and What Customers Need to Know

The market Fin is competing in has gotten crowded quickly, and the rebrand is partly a competitive positioning move. Salesforce has Agentforce, which includes AI customer service agents with monitoring and governance tooling for large deployments. Zendesk has integrated AI agents across its platform at scale. Freshdesk, Kustomer, and a range of AI-native startups offer automated resolution at various price points.

What distinguishes Fin within that field is live deployment scale and real resolution data. Most competitors are still accumulating the kind of evidence Fin can now report: 8,000 enterprise customers, 2 million conversations per week, customer-specific case studies with concrete numbers. The "Intercom" name, in a competitive pitch, required overcoming 15 years of positioning as a customer messaging tool. "Fin" can enter a pitch as an AI-native customer agent platform without that history. Over a two to three year horizon, as the company acquires new customers who never knew Intercom, the clean positioning becomes more valuable. McCabe's gamble is that the short-term costs of the brand transition are worth the long-term benefit of competing as an AI company rather than a legacy SaaS platform that added AI.

For teams already using Intercom's products, the practical changes are less dramatic than the announcement suggests. The Intercom help desk platform continues operating under its existing name. Intercom 2, a complete rebuild of the platform that launched alongside the rebrand, is a product improvement, not a replacement for existing workflows. The human agent tooling didn't go away; it became the layer for the 33% of cases that Fin doesn't resolve.

Fin Operator is in early access for Pro-tier customers now. Teams on the Pro tier interested in reducing the operational overhead of managing their Fin deployment can request access now, with general availability expected in summer 2026. Teams with high query volumes and well-maintained knowledge bases tend to see strong economics. Teams with complex product-specific questions and frequently changing documentation tend to see lower resolution rates, and Operator's knowledge management capability is the most direct lever for improving that.

The bet McCabe has made is that customer service AI is heading toward a concentrated market, and that winning it requires being unambiguously an AI agent company. That's a coherent position for a company with 2 million weekly conversations at 67% resolution from 8,000 live enterprise deployments. Operator extends the bet one layer further: not just that Fin can replace the humans answering questions, but that AI can also handle the operations work of keeping Fin running.

Teams thinking through the governance and rollout questions that come with deploying AI agents in customer-facing roles will find relevant frameworks in the Enterprise AI guide on AIntelligenceHub, covering deployment sequencing, policy controls, and accountability structures that apply as this category matures.