CEO scaling in the AI era: lessons from HubSpot's Brian Halligan

Executive overview

Starting a company has never been easier; scaling one has never been harder. More companies are forming, competition is intensifying, and the noise is deafening. The CEO job is also changing fast — shorter planning cycles, more decisions, less optionality.

The core challenge for founders is not starting: it's building the people, culture, and decision-making systems that let a company scale without breaking.

The LOCKS framework for evaluating CEOs

What to look for when assessing whether a founder can scale:

• L — Lovable: Would a talented 28-year-old crawl across broken glass to work for this person?
• O — Obsessed: Deep, long-standing fixation on the problem — not someone who found it six months ago
• C — Chip on the shoulder: Most top founders carry a boulder, not just a chip
• K — Knowledgeable: Domain expertise, not just enthusiasm
• S — Student: Constantly learning — history, strategy, peers, everything

Five-tool CEOs (can code, sell, inspire, have taste, and cast vision) are rare but increasingly common; AI does not easily fill those gaps.

Hiring and building the exec team

C-level hiring has a high mortality rate — roughly 50% turnover within 18 months is normal.

• Adults-table CEOs (100+ employees) spend half their time recruiting and interviewing
• Blind references are underrated; ask "Would you enthusiastically rehire this person?" or "Top 1% of your employees?"
• Parker Conrad's hack: send candidates a board deck under NDA before a 30-minute interview — red flag if they're only complimentary
• Avoid the big-company hire (ex-Salesforce, Google, Microsoft): impedance mismatch kills near-100% of them
• McKinsey hires almost never work; by definition they're conservative, founders are skeptical of conventional wisdom
• Prefer spiky candidates with clear strengths over safe candidates with the fewest weaknesses
• Shrink the interview panel (HubSpot went from 8 to 4 interviewers)
• Build like the 2004 Red Sox: homegrown talent supplemented by a few experienced free agents
• Give internal candidates the edge when it's close — they're underrated, external candidates are over-polished

What CEOs must learn on the job

Skills founders rarely have at the start but must develop:

• Giving hard feedback — the single most common growth area; layering or replacing co-founders and early hires is especially uncomfortable
• BS detection — the org and everyone around you is always selling; develops over time
• Inspiring people — unnatural at first; becomes the dominant job at scale
• Peer groups accelerate all of these; misery loves company, and advice lands better from peers than coaches

How the CEO job changes as you scale

• Startup phase: 90% perspiration, 10% inspiration
• Scale-up phase: 90% inspiration, 10% perspiration
• Planning cycles have compressed from one year to three months
• Optionality now has a massive tax — decide faster, walk through one-way doors sooner
• Excessive focus on the second act too soon kills the first; stay on the beachhead
• CEOs get put on a pedestal — every offhand remark becomes an initiative; tag communications (FYI vs. discuss vs. action)
• Trust surface is a scaling limit; most CEOs trust too few people for too long

Halligan's operating principles

Eat a shit sandwich — don't nibble Rip the band-aid. Do layoffs once, not in three tranches. Bad news doesn't improve with age.

Next play Taken from Coach K: make the error, forget it, run the next play. Don't compound mistakes with overreaction.

Never waste a good crisis Most good things at HubSpot came from crises. Swing the pendulum hard on the fix. HubSpot's worst outage forced a complete rethink of software deployment — no serious outage since.

If you want to kill a plant, have two people water it DRI (Directly Responsible Individual) is non-negotiable at scale. Everything important happens cross-functionally; one person must own it with the authority to direct others.

EV > TV > MEV (add CV in front) Solve for: Customer Value → Enterprise Value → Team Value → My Value. Managers who optimise for their team's metrics at the expense of peers or customers are a scaling cancer. Make it explicit in performance reviews, praise it publicly, track it in employee NPS by department.

Companies have one centre of gravity HubSpot was employee-centric for years (number 1 on Glassdoor). Shifted deliberately to customer-centric: customer panels at every management meeting and board meeting, comp tied to retention and NPS rather than revenue, repetition until it stuck.

Silver bullets don't exist Inside, every scaling company is two steps forward, one step back. No single hire, investor, or launch saves you. When you're a founder CEO, there's no one to rescue you — not your VCs, not your parents.

The future of go-to-market

• Enterprise sales hasn't changed much yet; deployed engineers (glorified solutions consultants) are the main AI-era addition
• The top of the funnel is about to break: buyers start in ChatGPT or Gemini, not Google
• Websites matter less; AI avatars that know everything about your product and pricing matter more
• Human enterprise sales (trust between carbon-based life forms) will be the last white-collar job AI replaces
• Every knowledge worker will have a personal AI agent in 3–5 years; it will join meetings as a participant, not just take notes

Starting vs. scaling

• Starting: easier than ever — AWS, AI tooling, global distribution; the number of companies formed will mushroom
• Scaling: harder than ever — more competition, noisier markets, distribution is the hard part
• Founders must learn distribution from scratch; they didn't grow up doing it
• Work-life balance is rare among the CEOs who scale fastest; the founders Brian coaches are all-in, seven days a week
• Don't start unless you're deeply obsessed; it's full contact and there's no one to rescue you