All podcasts / Lenny's Podcast / Summary

How to be a CEO when AI breaks all the old playbooks | Sequoia CEO Coach Brian Halligan

2026-02-15 - 75 min - source - Read full transcript

Lenny Rachitsky (host)Brian Halligan

Key insights

Scaling from founder to CEO of a large org means the job becomes mostly hiring.

Halligan says CEOs of companies over 100 employees (his 'adults table') spend roughly half their time recruiting and interviewing, a shift most founders don't expect. He calls this transition, and the org design work that comes with it, the core difference between the kids table and the adults table.

ceo-scaling

Reference checks beat interviews for spotting who will actually work out, but only if the questions are sharp.

Halligan says CEOs overrate their own interviewing instincts and underrate high-quality blind references. He prefers scaled, concrete questions such as 'on a scale of 1 to 10, how likely are you to rehire this person' or 'were they in the top 1% of your employees,' because vague strengths-and-weaknesses questions let references dodge honesty.

hiring-and-talent

Interview panels systematically pick the safest candidate over the highest-ceiling one.

HubSpot found that when panels scored candidates 1 to 4, they tended to hire the person everyone rated a uniform 3 (fewest weaknesses) over a 'spikier' candidate who scored some 4s and some 2s. Switching to spikier hires, and shrinking interview panels from eight people to about four, improved HubSpot's hit rate.

hiring-and-talent

Hire slow, fire fast, because roughly half of executive hires fail within 18 months regardless of process.

Halligan estimates that at least half of C-level hires are gone within 18 months industry-wide, so speed of correction matters as much as upfront diligence. He also warns against 'big company hires' from firms like McKinsey, Salesforce, or Google whose polish masks a poor fit for founder-led, convention-challenging cultures.

hiring-and-talent

Successful founder CEOs share a five-trait profile Halligan calls LOCKS.

Lovable (would you crawl across broken glass to follow them), Obsession (deep, long-standing founder-market fit rather than a six-month-old idea), Chip on the shoulder, Knowledgeable about the domain, and Student (constantly learning, like an LLM). He does not believe this is fully learnable; some people can't become this.

founder-psychology

A single directly responsible individual (DRI) becomes essential once a company scales past the room where everyone already knows what's going on.

In startup mode, cross-functional work happens naturally because the whole team is co-located and aligned. At scale, departments silo, so someone needs explicit authority to direct people outside their own reporting line toward a shared outcome, or the effort fails, illustrated by 'if you want to kill a plant, have two people water it.'

management-frameworks

Leaders should be evaluated and paid on company and customer value, not just their own team's numbers.

Halligan describes HubSpot VPs who optimized for their department's metrics (e.g., a sales leader maximizing bookings regardless of downstream service load) at the expense of the whole. HubSpot posted 'CV > EV > TV > Me' (customer, enterprise, team, then self) and tied comp and quarterly employee NPS by department to catch and correct this pattern before it became unrecoverable.

management-frameworks

A serious crisis, handled right, produces durable structural improvement rather than just a patch.

After a bad, all-day HubSpot outage, the team overhauled how they deployed software and moved from marketing-software thinking (where downtime is tolerable) to CRM thinking (where downtime directly damages customers doing business). Halligan says the company hasn't had a serious outage since, and generally argues teams should overcorrect rather than nibble at a fix.

management-frameworks

AI is starting to invert the sales funnel: buyers now research inside chat assistants before ever visiting a company's website.

Halligan expects the top of the funnel to shift from Google search and website browsing to conversations inside tools like ChatGPT or Gemini, which already know a company's product, pricing, and competitors. He predicts websites will pivot to an all-knowing avatar for that first conversation, with AEO (answer engine optimization) replacing SEO as the discovery lever, and a sales rep's own 'SE avatar' joining every call.

ai-and-go-to-market

Enterprise sales will be one of the last white-collar functions AI replaces, because it runs on human trust.

Halligan contrasts software development, customer support, and legal, where he says AI is already working incredibly well inside HubSpot, with enterprise sales, where trust built between two people over time is still the deciding factor. He notes that despite the hype, go-to-market processes at the companies he advises look almost identical to what enterprise sales looked like in the 1990s, aside from renaming SEs as 'deployed engineers.'

ai-and-go-to-market

Faster execution cycles are compressing CEO decision timelines and eroding the value of keeping options open.

Halligan says planning cycles have shrunk from roughly a year to about three months because AI-assisted teams ship faster (work that took a year now takes two months). This puts pressure on CEOs to make higher-quality, faster one-way-door decisions, since delaying a decision to preserve optionality now carries a real opportunity cost.

ceo-scaling

A near-fatal accident forced Halligan to confront mortality and step down as HubSpot CEO.

After a snowmobile accident that left him unconscious at the bottom of a cliff with 20 broken bones, Halligan decided he no longer wanted to run an 8,000-person company day to day. He handed the CEO role to Yamini Rangan shortly after recovering, framing the decision as 'life's short, don't waste it.'

founder-psychology

Books referenced

Marketing Lessons from the Grateful Dead - David Meerman Scott and Brian Halligan - Halligan's own book, referenced when comparing the Grateful Dead's disintermediation and first-principles thinking to Silicon Valley startups

Media referenced

The American Revolution - show - A roughly 10-hour Ken Burns documentary Halligan called his favorite recent watch, drawing a parallel between the Continental Army's near-losses and the two-steps-forward-one-step-back nature of scaling a startup
Long Strange Trip - podcast - Halligan's own podcast where he interviews top CEOs about the job of being a CEO, named after the Grateful Dead documentary

Companies

HubSpot - Halligan co-founded and ran it as CEO for about 20 years before stepping down; source of most of the examples in the episode
Sequoia Capital - Halligan is now Sequoia's in-house CEO coach, running peer groups (the kids table and adults table) for portfolio CEOs
Boston Red Sox - Halligan is a part owner; used as an analogy for building teams with a mix of homegrown talent and a few paid veterans
Rippling - Parker Conrad's NDA-plus-board-memo interview hack is cited as a way to see how a candidate really thinks
MongoDB - Halligan cites CEO Dev Ittycheria's stat of two C-level departures per year over a decade as evidence of how much exec turnover is normal at scale
Clay - CEO Kareem is cited as a rare exception among founders in that he insists on work-life balance and takes weekends off
Shopify - Cited via CEO Tobi Lutke as an example of a company doing well despite mediocre employee-satisfaction scores, contra the idea that being liked is a CEO's job

Techniques and frameworks

LOCKS framework - Halligan's rubric for what makes a founder CEO succeed: Lovable, Obsession, Chip on the shoulder, Knowledgeable, and Student
Directly Responsible Individual (DRI) - One person, not a committee, must own a cross-functional outcome; illustrated with 'if you want to kill a plant, have two people water it'
EV > TV > Me - A prioritization formula (later expanded to CV > EV > TV > Me) pushing leaders to optimize for customer and company value over their own team's or their own success
Blind reference checks - Asking pointed, scaled questions like 'would you enthusiastically rehire this person' instead of open-ended strengths and weaknesses questions
Next play mentality - Borrowed from Duke basketball coach Mike Krzyzewski: acknowledge a mistake, then move on immediately rather than compounding it
Flash tags for internal communication - A system co-founder Dharmesh Shah built to tag messages as a directive, a discussion topic, or just an FYI, to stop offhand comments from becoming company-wide mandates

Summary

Brian Halligan, HubSpot's co-founder and CEO for roughly 20 years, is now Sequoia's in-house CEO coach, running peer groups for portfolio founders split into a "kids table" (under 100 employees) and an "adults table" (over 100). The bulk of the conversation is a tour through the tactical differences between those two phases: adults-table CEOs spend half their time recruiting, they get religious about having a single directly responsible individual for cross-functional work, and they learn (often the hard way) that roughly half of their executive hires won't survive 18 months no matter how careful the process. Halligan's hiring advice is consistently anti-intuitive: distrust polished big-company resumes, prefer spikier candidates with real weaknesses over uniformly "safe" ones, and replace open-ended reference calls with sharp, scaled questions like "would you enthusiastically rehire this person."

A recurring structural theme is that things that work at small scale actively break at large scale, and the fix is usually to overcorrect rather than nudge. HubSpot's worst outage led to a full rebuild of how they shipped software; drifting too employee-centric (Halligan was once Glassdoor's #1-rated CEO) led to a deliberate multi-year swing toward customer-centricity, complete with customer panels inside management meetings and board sessions. He frames this with two of his "halaganisms": "when you have to eat a sandwich, don't nibble" (borrowed from Google CFO Ruth Porat, on layoffs) and "never waste a good crisis."

On AI specifically, Halligan argues the visible effects inside HubSpot are uneven: software development, support, and legal are already transformed, but enterprise sales and go-to-market are largely unchanged, still running on the same trust-building playbook that's been standard since the 1990s. What he does expect to change soon is the top of the funnel: buyers are starting their research inside ChatGPT or Gemini rather than a company's website, which he thinks will push companies toward an "all-knowing avatar" as the new homepage experience and toward AEO (answer engine optimization) as the new SEO.

The episode closes on more personal territory. Halligan is candid that he doesn't believe great founder CEOs have work-life balance, describing founders as "full contact" and on call seven days a week, an intensity he says has only increased since his own HubSpot years. He also recounts a near-fatal snowmobile accident that left him with 20 broken bones and, more consequentially, a decision made at the bottom of the cliff to step down as CEO of an 8,000-person company he no longer enjoyed running day to day. He handed the role to Yamini Rangan shortly after recovering.

Notable Quotes

"The thing about being a founder CEO is there's no one there to rescue you. Your parents aren't going to rescue you. Your VC is not going to rescue you. That kind of hits you when you hit your first crisis." - Brian Halligan

"Starting a company has never been easier. Scaling one into a durable, high-impact organization has never been harder." - Brian Halligan

"If you want to kill a plant, have two people water it." - Brian Halligan

"I think the go-to-market's going to get turned on its head... they'll stay in there [ChatGPT/Gemini/Anthropic] and do lots more research and be incredibly well educated. So your website's a lot less important." - Brian Halligan

"Life's short. Don't waste it." - Brian Halligan