How Intercom Defied the ‘SaaS Is Dead’ Narrative with an AI‑First Turnaround

1. Numbers Speak

Intercom now reports $400 M total ARR, with the AI Agent product Fin approaching $100 M. Fin’s growth rate is roughly four times the company’s overall rate, and the company has doubled its revenue growth for two consecutive years. The CEO expects Fin to represent half of revenue by early next year.

2. Why Intercom Had an Edge

Desperation : Growth had fallen well below industry averages, leaving little to lose and removing the inertia that protects “good enough” businesses.

CEO Change : Founder‑CEO Eoghan McCabe returned after a two‑year hiatus, finding a decision‑making process that was overly democratic and ineffective. He argued that new leadership could more easily discard the old approach.

Right Lane : Customer‑service is a natural landing zone for AI agents. Intercom already had an AI team before ChatGPT, and within six weeks of ChatGPT’s release they produced the first Fin prototype.

3. Creative Destruction – What Intercom Did

Ripped up old values and hired/fired around new ones

Rewrote the mission to focus on AI agents

Shifted performance targets to Fin revenue

Replaced the board’s “mature” members with startup founders

Allocated ~80% of R&D to Fin while it was still a single‑digit revenue share

Created a new brand and $1 M .ai domain, directing 100% of paid traffic there

Made all marketing, podcasts, and events about Fin, not Intercom

Chose risky brand positioning over incremental performance marketing

Expanded the AI team from ~6 to ~60 PhD‑level scientists, compensating them above all other teams

Built proprietary AI infrastructure and trained its own foundational models

Announced a $100 M cash‑reserve spend on AI when many still called it hype

Moved the entire customer base to a new, outcome‑based pricing model, voluntarily cutting about $60 M ARR

4. Organizational Reshaping

The first action after the CEO’s return was to rewrite company values and embed a quarterly scorecard that measured both performance and alignment with those values. Employees scoring below a threshold were respectfully let go, resulting in roughly 40% turnover over several quarters.

The board was also refreshed, swapping out “mature, experienced” directors for entrepreneurial ones, eliminating any internal veto power that could protect the legacy business.

4.1 Culture Before Business

New values were hard‑coded into internal systems, removing discretionary managerial decisions. The ongoing “soft coup” forced out dissenters and aligned the organization around the AI‑first vision.

4.2 Product Re‑engineering

Intercom’s pre‑AI product, the Resolution Bot, relied on handcrafted rules and keyword matching, which failed when user phrasing changed. The lesson cited was that “incremental improvement on a wrong paradigm is futile.”

Fin’s AI Engine™ is a multi‑stage Retrieval‑Augmented Generation pipeline optimized for support scenarios: minute‑level data refresh, proprietary retrieval and reranking models, and an application layer that can execute multi‑step business actions (e.g., refunds, order changes) through Fin Tasks.

Intercom also launched a no‑code “analyze‑train‑test‑deploy” loop for non‑technical business users, solving the “last‑mile” AI adoption problem and contrasting sharply with competitors that require months of engineering effort.

4.3 Pricing Overhaul

The old seat‑based pricing was described as “the most hated in SaaS.” Fin introduced a $0.99 per successful resolution model, charging only when the AI solved a ticket. Early on the model lost money because inference costs exceeded the price, but the CPO explained that short‑term loss was intentional to educate the market.

Switching the entire customer base to this model erased about $60 M of ARR, making the new pricing irreversible because customers’ purchasing logic had shifted.

4.4 Speed as Survival

The CTO announced a “2×” productivity goal, measuring monthly merged PR count. Actions included migrating the front‑end from Ember.js to React (because AI writes React more efficiently), allowing designers to ship code directly, and deploying an AI‑driven code‑submission agent.

The CPO dismantled ten‑year product processes: the classic triad (PM + designer + engineer manager), six‑week/‑year planning cycles, and static product team structures. Instead, staff were pooled into dynamic workstreams with two‑month planning horizons.

5. Independent Business Unit

Intercom created a separate brand, traffic channel, AI team reporting line, performance metrics, and pricing model—exactly the “independent business unit” prescribed by Christensen. By allocating 80% of resources to Fin while it was still a minor revenue source, the CEO bypassed the usual resource‑allocation constraints.

Because the CEO held ultimate authority, he could replace board members, fire 40% of staff, and lock in the irreversible decisions, ensuring the independent unit could not be re‑absorbed.

6. Lessons for SaaS

The market’s “SaaSpocalypse” narrative assumes SaaS products are static and will be displaced by AI agents. Intercom shows that the real threat is complacency; the companies that survive are those that become AI‑first and make irreversible bets.

“In a world of agent abundance, the workflow tools sold by our industry are clearly far less important… if you can’t become an agent company, your CRUD app business has a diminishing future.”

Forrester forecasts global SaaS spend to grow to $512 B by 2028, indicating that spend is not disappearing but being reallocated toward AI‑enabled services.

a16z and Hamilton Helmer argue that a software company’s moat is “process power” – deep understanding of work‑flow – not code. Intercom’s 40 M+ resolved tickets over years fed the Fin engine, turning it from a chatbot into an AI agent that embeds industry workflow knowledge.

Key Insights

Intercom’s X‑shaped growth curve stems from creative destruction, not just AI; the turning point was the CEO’s return.

Desperation can be an advantage—companies with modest growth become inert, while Intercom’s dire situation enabled bold moves.

Irreversible actions (e.g., cutting $60 M ARR, board overhaul, 80% R&D focus) distinguish true transformation from superficial AI add‑ons.

Only leaders with the authority to lock in those irreversible decisions can sustain an independent AI‑first unit.