Eoghan McCabe (founder and CEO)21 August 2025

How Intercom rose from the ashes by betting everything on AI

5Frameworks

14Insights

Episode overview

Eoghan McCabe explains how he returned to a slowing Intercom, concentrated the company on customer service, rebuilt its culture, and committed heavily to AI after GPT-3.5 appeared. A prototype of Fin emerged within six weeks, and its outcome-based pricing and rapid growth became the engine of Intercom's turnaround. The conversation also examines founder-led decision-making, workforce change, AI-driven job disruption, product-leadership development, therapy, and the human value that remains when automated output becomes abundant.

Key ideas

Intercom responded to near-negative growth by narrowing its strategy to customer service and making Fin central to the company's future.
An existing AI team and customer-data base allowed Intercom to build a working Fin prototype roughly six weeks after GPT-3.5 launched.
Fin charges per successfully resolved customer problem, aligning revenue with outcomes rather than seats or model-compute costs.
McCabe rebuilt the company through unilateral strategic choices, stricter values, quarterly performance scoring, cost cuts, and roughly 40% employee turnover.
Established companies cannot treat AI as a superficial feature; competing with AI-native startups requires expert talent, pervasive AI use, and unusually high commitment.
McCabe expects agents to automate repetitive operational work across customer experience, sales, accounting, legal work, and internal administration while collaborating with humans.
Intercom's product-leader alumni benefited from autonomous mini-CEO roles, first-principles reasoning, extensive framework-building, and the recruitment of founder-like employees.
McCabe credits long startup experience, sustained therapy with Yosi Amram, and surviving illness and professional setbacks with making him a more grounded leader.

Transcript available · source text is retained privately and is not published

Frameworks in this episode

Strategy4 steps

First-Principles Framework Method

Decompose every initiative into goal, mechanism, and value questions to build joined-up strategy — and mini-CEO operators.

Read →

Sales4 steps

Simpler-Fairer Pricing Reset

Deliberately give away revenue to simplify pricing — it lifts retention and restores a customer-obsessed culture.

Read →

Finance4 steps

Value-Anchored Outcome Pricing

Price the outcome, not the seat; set the number from customer value and let cost be your problem.

Read →

Leadership4 steps

Values as a Sharp Knife

Rewrite company values as deliberate filters, then hardcode a scoring formula that removes people who don't fit.

Read →

Leadership4 steps

Wartime Company Reset

When a platform shift threatens your core business, declare wartime and go all-in rather than hedge.

Read →

People & resources mentioned

Attributed to the moment in the episode. Timestamps are approximate.

People · 10

Eoghan McCabeMentionsIntercom founder and CEO who returned to lead its transformation into an AI-first company centered on Fin.
“With that, I bring you Owen McCabe.”
▶ Watch at ~2:30approx.
Des TraynorMentionsIntercom co-founder who alerted McCabe to the AI team's early product and is writing a book about AI disruption.
“So I got a text from Dez, my co-founder a week or so after the launch of GPT3.5”
▶ Watch at ~11:00approx.
Marc BenioffMentionsPreviously appeared on the podcast and discussed Agentforce and the future of sales and customer-experience jobs.
“we had Mark Beni off on the podcast he's you know all agent force agent force agent force”
▶ Watch at ~45:00approx.
Lenny RachitskyMentionstoday we've got another very special compilation episode something I've been pulling on more and more with the podcast and the newsletter
“they'll still want a Lenny that his has his own story and his own take and opinions and is a little imperfect”
▶ Watch at ~49:30approx.
Des TraynorMentionsIntercom co-founder who alerted McCabe to the AI team's early product and is writing a book about AI disruption.
“my co-founder Dez is writing a book at the moment and that's core to the idea here.”
▶ Watch at ~50:30approx.
Yosi AmramRecommendsCEO therapist and coach who worked weekly with McCabe for roughly twelve years.
“his name is Yosi Amram. Amazing guy.”
▶ Watch at ~59:00approx.
Ram DassMentionsSpiritual teacher cited in a discussion about ego and spiritual practice.
“there was you know Ram Das is this great spiritual teacher who died a few years ago”
▶ Watch at ~1:03:00approx.
Daniel KahnemanMentionshis influence was so huge you know and his intellect was so huge and his Insight was was so change making
“Reminds me of Daniel Conorman who wrote all these books about biases that we have”
▶ Watch at ~1:04:00approx.
Paul AdamsCoinedIntercom's chief product officer and author of a book about AI and customer service.
“Paul Adams our chief product officer I didn't even plan to show this he made this book recently the AI age”
▶ Watch at ~1:07:30approx.
Danny BoyleCoinedFilmmaker identified as the director behind 28 Years Later and earlier films discussed by the guest.
“the last movie I watched was 28 Years Later and that's by Danny Bole.”
▶ Watch at ~1:15:00approx.

Resources · 32

IntercomMentionscompany · Intercom
“The story of Intercom's transformation into Finn is incredible.”
▶ Watch at ~1:30approx.
FinUsesproduct · Intercom
“Finn, which is our AI agent, which is, you know, the future of the business”
▶ Watch at ~6:00approx.
G2Mentionswebsite · G2
“We're rated number one in G2.”
▶ Watch at ~6:30approx.
ChatGPTMentionssoftware · OpenAI
“I went back and one month later chat GBT was announced.”
▶ Watch at ~9:00approx.
GPT-3.5Usessoftware · OpenAI
“when GPT3.5 came out, they said, "This is different."”
▶ Watch at ~10:30approx.
WaymoMentionscompany · Alphabet
“Look at Whimo. So, Whimo doesn't crash. It has 3 5 times less crashes than humans.”
▶ Watch at ~15:00approx.
UberMentionscompany · Uber
“And it's really interesting to see Uber now offer women the option to call only female drivers.”
▶ Watch at ~15:30approx.
TwitterUseswebsite · Twitter
“I once had a Twitter poll or a survey on my newsletter”
▶ Watch at ~16:00approx.
GoogleMentionscompany · Google
“There are big companies like Google that do that.”
▶ Watch at ~34:30approx.
CoinbaseMentionscompany · Brian Armstrong and Fred Ehrsam
“they're not going to be happy in a company like Intercom or Coinbase or any of these strong organizations.”
▶ Watch at ~35:30approx.
SlackUsessoftware · Slack Technologies
“Who wants to go to war every day with your organization and in Slack?”
▶ Watch at ~35:30approx.
ChatGPTUsessoftware · OpenAI
“anytime they were they had a question for anyone else on the team they first asked chat GBT about it”
▶ Watch at ~55:00approx.
PerplexityMentionssoftware · Perplexity AI
“I did a interview with the Perplexity founders.”
▶ Watch at ~55:00approx.
The AI Age and the Transformation of Customer ServiceCoinedbook · Paul Adams
“he made this book recently the AI age and the transformation of customer service”
▶ Watch at ~1:07:30approx.
Jobs to Be DoneMentionsother · Bob Moesta and Clayton Christensen
“You guys popularized jobs to be done.”
▶ Watch at ~1:10:30approx.link ↗
RICECoinedother · Intercom
“you guys invented rice.”
▶ Watch at ~1:10:30approx.
IBMMentionscompany
“If if I told you that IBM had made the most wildly innovative coding assistant, you'd find it hard to believe.”
▶ Watch at ~1:12:00approx.
FinRecommendsproduct · Intercom
“So go check out Finn, please. Finn.ai.”
▶ Watch at ~1:12:30approx.
Nuclear War: A ScenarioRecommendsbook · Annie Jacobsen
“the most recent book I read is a book called Nuclear War Scenario”
▶ Watch at ~1:14:00approx.
True DetectiveRecommendstv show · Nic Pizzolatto
“the first and last TV show I loved was True Detective 1. It was just like incredible.”
▶ Watch at ~1:14:30approx.
TrainspottingMentionsfilm · Danny Boyle
“grew up with you know you know train spotting was 28 days later then he made a movie called Sunshine.”
▶ Watch at ~1:15:00approx.
28 Days LaterMentionsfilm · Danny Boyle
“grew up with you know you know train spotting was 28 days later then he made a movie called Sunshine.”
▶ Watch at ~1:15:00approx.
SunshineMentionsfilm · Danny Boyle
“then he made a movie called Sunshine.”
▶ Watch at ~1:15:00approx.
28 Years LaterRecommendsfilm · Danny Boyle
“but I love 28 years later.”
▶ Watch at ~1:15:30approx.
FellowRecommendscompany
“I've been buying products by Fellow. They're remarkably good for consumer products like different.”
▶ Watch at ~1:16:00approx.
Porsche 911Recommendsproduct · Porsche
“I bought a Porsche 911 recently and that is a beautiful product.”
▶ Watch at ~1:16:30approx.
SavorableCoinedsoftware · Lenny Rachitsky
“Fun fact, I built an app once called savorable that helped you savor the moment.”
▶ Watch at ~1:19:00approx.
QwitterCoinedwebsite · Eoghan McCabe
“Speaking of apps, I was doing research on you in preparation for this and I didn't realize you built Twitter back in the day.”
▶ Watch at ~1:19:30approx.
FlippaUseswebsite
“I think there's a website called Flip It where you could sell websites.”
▶ Watch at ~1:20:00approx.
Apple PodcastsMentionssoftware · Apple Inc.
“you can subscribe to the show on Apple Podcasts, Spotify, or your favorite podcast app.”
▶ Watch at ~1:22:30approx.
SpotifyMentionssoftware · Spotify
“you can subscribe to the show on Apple Podcasts, Spotify, or your favorite podcast app.”
▶ Watch at ~1:22:30approx.
Lenny's Podcast websiteRecommendswebsite · Lenny Rachitsky
“You can find all past episodes or learn more about the show at lennispodcast.com.”
▶ Watch at ~1:23:00approx.

Spot an error or want something removed? Request a correction or removal.

Insights & moments

The myth-busts, hot takes, explainers, and tools worth keeping.

Hot Take· 6

Hot Take12:30

Why an AI agent is more personal than a human

Intercom's founding mission was 'make internet business personal,' and McCabe wrestled with whether that survives an AI pivot. He argues it does: an instantly available, expert, consistent, friendly agent for every customer at every hour is more personal than making people wait days for a canned response. He concedes it may be post-rationalization but says AI is genuinely superior at the things we call personal and human.

Intercom's mission was 'make internet business personal'
McCabe questioned whether that mission still made sense with AI
An always-available, consistent, expert agent beats a multi-day canned reply
He argues AI can be superior at qualities we describe as personal and human

“so much more personal than making them wait two 3 4 days for a crappy canned response.”
Eoghan McCabe · 14:00

“it actually is superior at the things we describe as personal and human relative to humans themselves.”
Eoghan McCabe · 14:00

#ai#customer-service#philosophy#personalization

Hot Take15:00

AI is often superior — the Waymo case

McCabe uses Waymo to argue AI often beats humans on practical dimensions: it crashes far less, doesn't bug you, has no hygiene problems, and doesn't take wrong turns. He notes Uber now offers women female-only drivers because women feel safer in Waymos. He is pro-human for meaning and creativity, but wants AI and robotics for the practical, efficient 'glue' between the human parts of life.

Waymo crashes 3-5x less than humans, per McCabe
AI avoids human annoyances: hygiene issues, wrong turns, unwanted chat
Uber offering female-only drivers reflects women feeling safer in Waymos
McCabe wants humans for meaning, AI for practical, efficient value

“So, Whimo doesn't crash. It has 3 5 times less crashes than humans.”
Eoghan McCabe · 15:00

“women love Whimo because they feel safer.”
Eoghan McCabe · 15:30

#ai#waymo#automation#hot-take

Hot Take26:00

Founder mode: a soft coup, 40% turnover, 98% approval

McCabe describes taking an authoritarian, top-down approach: cutting costs, picking one lane (service), betting ~$100M of cash on AI, and rewriting the values as 'a sharp knife' to cut out ineffective parts of the company. He hardcoded a formula scoring people on both goals and value-behavior. It triggered a soft coup and letters to the board, ~40% turnover, but ended with a 98-99% approval of leadership and strategy.

Cut costs aggressively and killed projects while nearing negative growth
Picked one lane — service — over keeping an $80M ARR side business
Committed nearly $100M of Intercom's own cash to AI
Rewrote values as 'a sharp knife' and hardcoded a performance+behavior formula
Provoked a soft coup and ~40% turnover, but reached 98-99% leadership approval

“I rewrote the values designed to be a sharp knife to cut out the parts of the company that I just knew wouldn't be effective.”
Eoghan McCabe · 28:30

“it is the responsibility of the CEO to make brave and hard decisions unilaterally.”
Eoghan McCabe · 31:30

#founder-mode#leadership#culture#turnaround

Hot Take36:00

The unglamorous reality of being a CEO

Asked if he regretted returning, McCabe says never — it was cathartic — but he's honest that most days aren't the swashbuckling adventure people imagine. Success means reviewing bonus policy and comp proposals, accountability meetings, 8-10 meetings a day, and racing through emails so no one feels slighted. He calls corporate life something that 'kind of sucks,' especially for high-agency people like him, and stays only for a broader mission.

He never regretted returning; it was deeply cathartic after running away while sick
Most CEO days are bonus policy, comp proposals and accountability meetings, not bold moves
Rushing from meeting to meeting, 8-10 a day, isn't how he thinks life should be lived
He only stays because a broader mission makes it worthwhile

“It'll be rushing from meeting to meeting having 8 n 10 meetings a day. I don't happen to believe that that's a great way to…”
Eoghan McCabe · 38:00

“It's not it's corporate life kind of sucks, particularly for people like me.”
Eoghan McCabe · 38:30

#leadership#ceo#founder#reality

Hot Take45:30

Jobs will go — and it's the demeaning work that should

McCabe doesn't find job loss touchy: technology has always taken repetitive, demeaning, dangerous jobs, while population, GDP and longevity rose and crime fell. He points to the horror of a human clicking a macro to answer the same question all day, and to the bright spark in workers who'd flourish doing something else. He expects fewer people needed in CX and repetitive sales/SDR roles, mildly disagreeing with Marc Benioff.

Technology has always removed repetitive, demeaning, dangerous jobs
Historically, population, GDP and longevity rose while crime fell alongside that
Answering the same question all day by clicking a macro is a 'horrible use of a human life'
Expects fewer people in CX and repetitive sales/SDR roles
Slightly disagrees with Benioff, who said sales jobs aren't going away

“technology has done a really good job at stealing jobs that were repetitive, demeaning, dangerous.”
Eoghan McCabe · 45:30

“Like what a horrible use of a human life.”
Eoghan McCabe · 47:30

#ai#jobs#automation#future-of-work

Hot Take49:00

Why people will still pay more for humans

Even as AI aggregates and creates content, McCabe argues efficiency is not what humans value most — if it were, we'd always buy the cheapest of everything. People value beauty, human stories, heart and connection, and will pay more for an imperfect human like Lenny with his own take. The abundance of AI makes automated things worth zero, which is exactly why quality, craft, art and humanity command a premium.

Efficiency is not the top thing humans value — or we'd always buy the cheapest option
People value beauty, human stories, heart and connection
They'll pay more for an imperfect human with their own story and opinions
AI abundance drives the value of automated output toward zero
Quality, craft, art and humanity are what command a premium

“humans value things like beauty and human stories and human heart and connection.”
Eoghan McCabe · 49:30

“The abundance of AI is going to make automated things worth zero.”
Eoghan McCabe · 50:00

#ai#human-value#content#economics

Explainer· 2

Explainer23:00

The 99-cent bet: price on value, not cost

Finn charges 99 cents per resolved ticket — outcome-based pricing aligned to value delivered. McCabe explains that although research showed businesses spent $20-30 per ticket resolution, people simply won't value digital work as much as human work, so they landed on 99 cents as the nexus of maximum earnings and maximum palatability. Early on the cost was higher than the price, but he bet costs would fall — and they did.

Finn charges 99 cents per resolved customer ticket
Research showed many businesses spent $20-30 per ticket resolution
People undervalue digital work versus human work, even when it's better
Early on it cost more than 99 cents to resolve a ticket; McCabe bet costs would drop
His principle: pricing should come from value, not cost — cost is the company's problem

“if someone is not prepared to pay 99 cent for us to rapidly and elegantly, perfectly and excellently solve their customer's problem. We need to…”
Eoghan McCabe · 25:00

“I always believe that value should that pricing should come from value and not from cost. Cost is our problem.”
Eoghan McCabe · 25:00

#pricing#outcome-based-pricing#finn#unit-economics

Explainer1:04:30

Why Intercom produces so many top product leaders

Told that Intercom ranks number one for producing product leaders (alongside Palantir and Stripe), McCabe points to a deeply producty culture, a sprawling strategy that gave PMs mini-CEO autonomy, and a first-principles methodology where they built frameworks for everything. They also hired founder types — good founders, often not great employees — with a pitch to learn how great companies are built and then go start their own.

Intercom ranked number one for producing future product leaders, near Palantir and Stripe
A deeply producty culture attracted and retained strong product people
A sprawling strategy gave PMs real autonomy to act like mini-CEOs
A first-principles methodology meant building frameworks for everything (RICE, jobs-to-be-done)
They hired 'founder types' who were great founders but often not great long-term employees

“kind of the product of our big messy strategy was that we had PMs that got to act like mini CEOs.”
Eoghan McCabe · 1:06:00

“we would create frameworks for everything.”
Eoghan McCabe · 1:06:30

#product-management#culture#hiring#first-principles#intercom

Story· 4

Story05:30

Finn's growth: 1 to 12M ARR in year one, north of 300%

Eoghan McCabe lays out how well Intercom's AI agent Finn is performing. Intercom now ranks in the 15th percentile of ~120 public B2B software companies for ARR growth, and Finn itself is growing north of 300%, going from $1M to $12M ARR in its first year and heading past $100M ARR. In the customer-experience agent category, Intercom claims it is biggest by customer count and revenue and rated number one on G2.

Intercom sits in the 15th percentile of ~120 public B2B software companies for ARR growth
Finn is growing north of 300%
Finn grew from $1M to $12M ARR in its first year, now in solid mid-8-figure ARR
On track to pass $100M ARR with Finn in under three quarters
Claims biggest by customer count and revenue in the CX agent category, number one on G2

“The first year it grew from 1 to 12 million air or we're now in solid mid 8 or growth there.”
Eoghan McCabe · 06:00

“We win all our head-to-heads and our direct competitor bake offs. We're rated number one in G2.”
Eoghan McCabe · 06:30

#intercom#finn#ai-agents#growth#saas

Story07:30

Five quarters of decline and the pride that pulled him back

McCabe describes the state of Intercom before the AI pivot. After he left the CEO role in 2020 while sick with mold toxins and a tick-borne illness, the company grew bloated and unfocused, with revenue growth in the low single digits. Intercom saw five successive quarters of declining net new ARR and was about to hit zero, meaning negative growth. What pulled him back was pride and an unwillingness to watch the company fade.

McCabe left the CEO role in 2020 while sick with mold toxins and a later tick bite
The company became bloated, with a diluted, unfocused strategy
Revenue growth fell to the low single-digit percent
Five successive quarters of sequential decline in net new ARR, about to hit zero
Pride, not just the AI opportunity, drove his return

“we had five quarters of successive, you know, sequential decline in our net new error.”
Eoghan McCabe · 08:30

“still had a lot of pride for this damn thing and didn't want to see it end in a way that was so different from…”
Eoghan McCabe · 08:30

#intercom#turnaround#founder#decline#leadership

Story10:00

A Finn prototype in six weeks — and nothing to lose

McCabe credits luck: Intercom already had AI engineers and a customer-service ML team when GPT-3.5 launched. Just six weeks after launch they had a working beta of Finn, tapping a base of 30,000 paying customers and billions of data points. He is candid that the 'brave maverick' framing is partly retrospective — the company was in real trouble and had nothing to lose.

Intercom already had AI engineers and a customer-service ML team when GPT-3.5 launched
A working Finn beta existed just six weeks after GPT-3.5's launch
They had ~30,000 paying customers and billions of data points to work with
McCabe frames the bold move as partly born of desperation, not just courage

“he said the AI team have something interesting and they actually think we could make a product out of this”
Eoghan McCabe · 11:00

“I I don't know a a single company of our size and age that has pivoted this hard to AI and being as successful as…”
Eoghan McCabe · 11:30

#intercom#finn#gpt-3.5#product#pivot

Story58:00

Therapy, ego death, and becoming a better leader

Asked about a visible personal transformation, McCabe credits three things: 14 brutal years in a startup, 12 years of weekly therapy with a CEO-specialist coach (Yosi Amram), and a two-year period away, sick and attacked in the press, that eviscerated his ego. He argues great therapy smooths away only the counterproductive edges — the insecurity and triggers — and that the identity of being a perfect, brilliant leader had to die because it's so limiting.

Three sources: 14 years in a startup, 12 years of therapy, and two humbling years away
His coach Yosi Amram only worked with CEOs and helped him know himself
Good therapy removes only the counterproductive edges — insecurity, triggers
Understanding yourself lets you understand and communicate with others better
The ego-identity of being a perfect brilliant leader had to die because it's limiting

“the interesting thing about 12 years of weekly therapy and spiritual work is that it takes your edges off. But they are all edges that…”
Eoghan McCabe · 1:00:00

“your image that you are this perfect brilliant leader that all successful founders form when they are successful had to die.”
Eoghan McCabe · 1:01:30

#therapy#ego#personal-growth#leadership#founder

Takeaway· 2

Takeaway16:30

Giving away $50M in ARR to simplify pricing

Intercom's old pricing was infamous — a meme, mocked in viral tweets — because an unfocused 'all things for all people' strategy forced complex, multi-metric pricing. On his return, McCabe deliberately wrote down roughly $50M in ARR to give customers far simpler, fairer pricing, betting that predictability makes customers stay longer and builds a healthier relationship. He tied it to a new 'customer obsessed' value.

Intercom's old pricing was widely mocked, even in viral memes
An unfocused strategy forced complex multi-metric, multi-tier pricing
McCabe gave away roughly $50M in ARR to simplify and lower prices
Simpler, fairer, predictable pricing makes customers stick around longer
Complex unfair pricing eroded the internal belief that the company cared about customers

“we actually have already given away something like something like $50 million in AORA.”
Eoghan McCabe · 18:00

“when people feel like they have far simpler, more predictable, fairer pricing, they'll stick around longer”
Eoghan McCabe · 18:30

#pricing#strategy#saas#customer-obsession

Takeaway50:30

You don't have a choice — and if you can't, hire a kid

McCabe's core advice: with AI you don't have a choice, because it will disrupt every category in the most aggressive ways, and if you're not in it you'll be kicked out of all of it. Winning requires real AI talent (young and senior) and brutal hours — young AI companies work 12 hours a day, 365 days a year. If a previous-generation founder won't roll up their sleeves and work that hard, his blunt advice is to become chairperson and hire a kid.

AI will disrupt all categories, aggressively — 'you don't have a choice'
Bring in real AI scientists and leaders, plus young talent
Young AI companies work ~12 hours a day, 365 days a year, and vibe-code by default
Don't half-ass it by sprinkling a little AI onto the same culture
If you won't work that hard yourself, become chairperson and hire a kid

“if you're not in it, you're about to get kicked out of all of it.”
Eoghan McCabe · 51:30

“Hire a kid cuz you're in the wrong job, buddy.”
Eoghan McCabe · 53:30

#ai#advice#talent#disruption#hard-work

Key ideas

Frameworks in this episode

First-Principles Framework Method

Simpler-Fairer Pricing Reset

Value-Anchored Outcome Pricing

Values as a Sharp Knife

Wartime Company Reset

People & resources mentioned

People · 10

Resources · 32

Insights & moments

◆Hot Take· 6

Why an AI agent is more personal than a human

AI is often superior — the Waymo case

Founder mode: a soft coup, 40% turnover, 98% approval

The unglamorous reality of being a CEO

Jobs will go — and it's the demeaning work that should

Why people will still pay more for humans

✶Explainer· 2

The 99-cent bet: price on value, not cost

Why Intercom produces so many top product leaders

❝Story· 4

Finn's growth: 1 to 12M ARR in year one, north of 300%

Five quarters of decline and the pride that pulled him back

A Finn prototype in six weeks — and nothing to lose

Therapy, ego death, and becoming a better leader

▲Takeaway· 2

Giving away $50M in ARR to simplify pricing

You don't have a choice — and if you can't, hire a kid

Hot Take· 6

Explainer· 2

Story· 4

Takeaway· 2