How to Be a CEO When AI Breaks All the Old Playbooks

Starting a company has never been easier. Scaling one into a durable, high-impact organization has never been harder. That tension sits at the center of everything Brian Halligan thinks about as Sequoia’s in-house CEO coach.

The View from 200 CEOs

Brian Halligan co-founded HubSpot, ran it as CEO for about 15 years, and stepped down after a near-fatal surfing accident rewired his priorities. Now he coaches Sequoia’s fastest-growing founders, runs CEO dinners, and hosts a podcast called Long Strange Trip. The vantage point is unique: he sees patterns across dozens of companies simultaneously, watching the same challenges surface in different forms.

His coaching philosophy starts with a refusal to prescribe. He doesn’t tell CEOs what to do. Instead, he tries to understand them deeply, helps them see blind spots, and makes suggestions they can accept or reject. His model: a caddy in golf, not a swing coach.

“I feel like I have a cheat code in life being a CEO coach, because I learn from each one of them and I stitch together these patterns.”

The thing he watches most closely is what he calls “constructive dissatisfaction,” a restlessness that prevents founders from ever feeling satisfied, even when things are going well. The best CEOs are never done iterating.

The LOCKS Framework

Halligan evaluates founders through four dimensions he calls LOCKS:

L - Likability. Not charm or charisma, but the ability to attract and retain talent. The best CEOs have a magnetic quality that makes A-players want to work for them. Halligan notes this has become more important over the last decade as the war for talent has intensified.

O - Obsession. Founders need to be genuinely obsessed with their problem space. Not just interested, not just committed, but unable to stop thinking about it. This is the trait most correlated with surviving the inevitable crises.

C - Communication. The ability to articulate vision clearly, both internally and externally. Halligan specifically looks for CEOs who can explain complex ideas simply, because that skill scales.

K - Knowledge. Deep domain expertise in their market. The best founders know their space better than anyone else in the room.

S - Speed. The ability to make decisions quickly and move fast. In the AI era, this matters more than ever because the optionality cost of moving slowly has skyrocketed.

He acknowledges the framework has evolved. Speed was always important, but AI has made it the single biggest differentiator. CEOs who can compress decision cycles from weeks to days now have a structural advantage.

Hiring: The 2004 Red Sox Theory

Halligan’s most provocative hiring insight comes from baseball. The 2004 Boston Red Sox, famous for breaking the 86-year championship drought, were not a team of the best individual players. They were a team of “idiots,” as they called themselves: high-character, gritty, team-first players who complemented each other.

His hiring framework:

Hire “spicy” candidates over consensus picks. When an interview panel rates everyone 7/10, that’s a warning sign. You end up with a team of B-players. Instead, look for candidates who generate strong reactions: some interviewers give them a 9, others a 4. Those polarizing candidates tend to be the ones who bring genuine differentiation.

“When everyone rates a candidate 7 out of 10, you’ve got a 7. When someone rates them a 9 and someone else a 4, dig in. That 9 might be seeing something everyone else is missing.”

Reference calls are where the real signal is. Halligan treats reference calls as the most important part of the hiring process, not a formality. His technique: ask the reference to compare the candidate to everyone else they’ve worked with in that role and place them in a percentile. “Top 5%? Top 20%? Average?” The pause before they answer tells you everything. He also asks, “If you were starting a company tomorrow, would you hire this person?” If they hesitate, that’s a no.

Homegrown talent often outperforms big-company hires. One of his most consistent observations across Sequoia’s portfolio: the people who grow up inside the company frequently outperform expensive external hires from Google, Meta, or McKinsey. The external hires bring process and credibility but often struggle with the ambiguity and speed of a startup. He recommends a mix but warns against over-indexing on pedigree.

Born or Made?

Halligan believes great CEOs are roughly 50% born, 50% made. The “born” part is temperament: risk tolerance, emotional resilience, the ability to handle ambiguity, and the obsessive drive. You can’t teach someone to care more.

The “made” part is everything learnable: giving feedback, running meetings, building culture, managing boards. These are skills that improve dramatically with coaching and deliberate practice. His observation is that most CEO failures come from fixable problems: not giving direct enough feedback, not firing fast enough, not delegating well.

The biggest coaching intervention he makes: helping CEOs give harder, more direct feedback. Most founders are naturally empathetic and struggle to deliver difficult messages. He teaches a framework: give feedback within 24 hours, be specific about what you observed, explain the impact, and offer a clear path forward.

AI Reshapes Go-to-Market

This is where Halligan’s CEO perspective intersects with the AI moment. He sees the go-to-market motion being completely rewritten:

SDRs are the first casualties. The traditional sales development representative role, cold-calling and emailing to generate leads, is being automated rapidly. AI agents can do this at scale with better personalization. Halligan estimates most SDR teams will be 80% smaller within two years.

Marketing is getting compressed. Content creation, SEO, email campaigns, ad targeting: every part of the marketing stack is being automated. The remaining human marketers will be strategists and brand-builders, not executors.

Enterprise sales is the last to fall. This is Halligan’s most interesting prediction. Complex enterprise deals involving multiple stakeholders, long sales cycles, relationship-building, and navigating organizational politics are the hardest for AI to replicate. The “human in the loop” matters most when the deal involves trust, nuance, and organizational change management.

Vertical SaaS is the big opportunity. Halligan is most bullish on AI-native vertical SaaS companies, software built for specific industries (healthcare, legal, construction) that deeply understands the domain and can automate entire workflows. He thinks horizontal SaaS will face more pressure because AI makes it easier for companies to build their own internal tools.

“Enterprise sales will be the last white-collar job AI replaces. When you’re selling a $2 million deal to a Fortune 500 company, the relationship and trust still matter enormously.”

Forward Deployed Engineers

One pattern Halligan has noticed across Sequoia’s portfolio: the rise of “forward deployed engineers” (FDEs). These are engineers who sit with customers during the sales and onboarding process, customizing the product in real-time.

The model, pioneered by companies like Palantir, is becoming more common because AI products often require significant configuration and integration. FDEs combine technical ability with customer empathy, and Halligan sees them as the bridge between product and market in AI-native companies.

This changes the traditional org structure. Instead of a clean separation between engineering, sales, and customer success, you get hybrid roles that blur the lines. CEOs need to figure out how to manage, compensate, and develop these people differently.

Halliganisms

Halligan has collected dozens of aphorisms over his career. A few that stood out:

“Culture is not ping-pong tables. Culture is who you hire, who you fire, and who you promote.” He’s direct about this: the signals that actually shape culture are personnel decisions, not perks.

“Your first 50 employees are your culture. After that, culture manages itself, for better or worse.” The implication: founders need to be obsessively intentional about early hires because those people will replicate themselves.

“Speed is the ultimate competitive advantage. Not strategy, not capital, not talent. Speed.” This has become his core thesis in the AI era. The founders who win are the ones who can make decisions and ship faster than everyone else.

“The best CEOs are readers, not just leaders.” He pushes every CEO he coaches to read voraciously, not just business books but history, biography, science. The meta-insight: diverse inputs lead to better pattern recognition.

“Hire people who have done it at scale, not just people who have ideas about doing it.” The distinction between operators and advisors. He values execution experience over theoretical knowledge.

The Surfing Accident and What It Taught Him

Halligan was nearly killed in a surfing accident several years ago, breaking 20 bones and ending up with 33 screws in his body. That experience fundamentally changed his relationship with time and priorities.

Before the accident, he was a classic founder-CEO: all-consuming, always on, 7 days a week. After, he became far more intentional about what he worked on and who he spent time with. He handed the CEO role at HubSpot to Yamini Rangan, who he calls one of the best operators he’s ever seen.

His advice from that experience: you don’t need a near-death experience to reprioritize. Most founders wait too long to step back and ask whether they’re spending their time on what actually matters.

A Few Observations

This conversation is less about AI itself and more about the humans navigating it. Halligan’s value isn’t in predicting which AI models will win. It’s in his pattern recognition across hundreds of founders dealing with the same moment simultaneously.

• The LOCKS framework is simple but useful precisely because it’s not about skills. It’s about temperament. You can teach a CEO to run a board meeting; you can’t teach them to be obsessed with their problem.
• The “spicy hire” concept deserves wider adoption. Consensus hiring is one of the most common failure modes at scaling companies, and few people articulate the alternative as clearly as Halligan does.
• His prediction about enterprise sales being the last white-collar job AI replaces feels counterintuitive but well-reasoned. The complexity isn’t in the selling; it’s in the organizational change management that accompanies any large purchase.
• The forward deployed engineer model is quietly reshaping how AI companies go to market. It’s worth watching whether this becomes the default or remains a specialty approach.
• There’s something refreshing about a CEO coach who openly says half of being a great CEO is innate. Most coaching narratives pretend everything is learnable. Halligan’s honesty about the limits of coaching makes his advice about the learnable parts more credible.