Context

Andrew Anagnost is CEO of Autodesk, the design and construction software company behind AutoCAD, whose tools are used to design most of the built and manufactured world. Host Jeff Berman traces his path from a teenage high school dropout to a Stanford PhD to Autodesk CEO, including a contentious co-CEO period before he won the sole role. This episode matters to PMs because Anagnost lays out a concrete, repeatable test for deciding which products or business lines to kill, a hop-versus-leap framework for adjacency decisions, and a leadership philosophy that treats empathy as the mechanism that makes courageous, unpopular decisions land without breaking trust.

The Big Idea

Leading a company through a hard pivot requires a simple, repeatable test for what to kill or keep, the discipline to fix unglamorous infrastructure before building anything new on top of it, and treating empathy not as softness but as the thing that makes courageous, unpopular decisions actually work.

Anagnost rebuilt Autodesk's back office and killed a large share of its product bets using a three-question test before making any of the bigger strategic moves, like large acquisitions, that reshaped the company. He frames courage without empathy as "just meanness," arguing that understanding how a hard decision lands with the people affected is what determines whether it succeeds.

Key Insights

1. Fix the boring back office before building the exciting new thing

• What: on his first day as sole CEO, Anagnost's first major move was shifting money and attention into Autodesk's back-office systems, the unglamorous infrastructure behind billing, subscriptions, and scale, rather than into new products.
• Why it matters: he explains the logic directly: "I could build the best products in the world right now and I couldn't get it out to the customer." A great product built on infrastructure that can't scale or sell effectively is not actually a working business.
• Example: Anagnost had previously led Autodesk's transition to a subscription model, so he already understood that the back office was the bottleneck preventing new products (including future AI products) from reaching customers at scale.

2. A three-question test decides what to kill

Anagnost uses a simple, repeatable test before deciding whether to keep or cut a product or initiative: does the company have a genuine right to win in this space, how adjacent is it to the current business (closer bets are easier to execute), and does the team have the right talent and product vision to pull it off. Applying this test early in his CEO tenure, he pulled a large number of internal projects he judged to be "weeds," by his own account getting the call wrong on at least a couple of them, but estimates he was right roughly 70% of the time, which he considers a strong success rate for this kind of decision.

3. Most adjacency moves should be hops, not leaps

Anagnost distinguishes between "hopping" into an adjacent market (a smaller, closer bet that's easier to walk back if wrong) and "leaping" into a distant one (a bigger bet that's much harder to execute or reverse). His rule: leap only when your core business faces a genuine paradigm shift, the way Netflix had to leap from DVDs to streaming, or the way Autodesk leapt early into AI as a core technology bet. Otherwise, hop into adjacencies close enough to the existing business that the company's current capabilities and "right to win" still apply; a leap taken too early burns excessive money, and one taken too late means falling behind with no time left to catch up.

4. Rebuilding trust after a leadership overhaul requires repeating the "why," not just the "what"

• What: roughly three-quarters of Autodesk's leadership team either left voluntarily or was asked to leave under Anagnost, following a bruising co-CEO period where the company was roughly split between supporters of him and of the outgoing leader.
• Why it matters: he had to earn trust from both factions simultaneously, in his words being "a whole company leader," inside a narrow window of about one to two years before people would start seriously questioning his direction.
• Example: his method was to over-communicate the "why" behind decisions repeatedly while leaving more of the "what" open for others to challenge and shape, and to deliberately cross-pollinate people from both factions into new roles rather than punishing anyone for which side of the co-CEO fight they had supported.

5. A co-CEO structure can stall a large company for months

For six months during Autodesk's leadership transition, the company operated under a co-CEO arrangement between Anagnost and a leader from the prior era, during which no major strategic decisions were made. Anagnost believes the board created this structure less as a genuine governance solution and more as "theater" to visibly justify a fair-seeming process while buying time to assess the situation, but he is candid that the arrangement "wasn't great" and effectively froze company decision-making for half a year. The broader lesson: a compromise structure designed to look fair to outside stakeholders can carry a real operational cost if it isn't paired with genuine, aligned partnership between the people sharing the role.

6. Vertical, precise tools beat horizontal ones for domain-specific work

Anagnost argues that using a large, general-purpose AI model to solve a narrow, high-precision technical problem is expensive and imprecise, because it requires burning significant compute on complex prompt chains to compensate for the model not already knowing the domain. A smaller, purpose-built model that already understands the specific problem needs far less prompting to reach the same answer, delivering better precision, speed, and total cost of ownership. He predicts this "expensive to use a horizontal tool to solve vertical problems" dynamic will define enterprise AI software strategy over the next year or two, favoring companies that invest in custom, domain-specific models over those relying purely on general-purpose ones.

Mental Models & Frameworks

Courage, scholarship, and empathy as one leadership system

• Courage: the willingness to attack your own successful business lines, tell people hard truths, and admit "I don't know," rather than protecting comfortable but weakening parts of the business.
• Scholarship: learning enough about a situation before acting so that boldness doesn't collapse into recklessness; Anagnost keeps a sign in his office reading "impulsiveness is not a virtue," distinguishing genuine decisiveness from acting without enough understanding.
• Empathy: understanding how a courageous decision will land with the specific people it affects, and communicating both why the decision is being made and what will be done to support people through it, not simply softening or delaying the decision itself.

Anagnost's framing is that courage without empathy is "just meanness," while empathy without courage collapses into sympathy or paralysis. Use this model whenever a leader has to make a decision (a layoff, a product cancellation, a strategy reversal) that will hurt some group inside the organization: pair the courage to act with the discipline to understand and communicate the real impact, rather than treating empathy as an obstacle to decisive leadership.

Hop versus leap for growth bets

A framework for deciding how far outside the current business to expand: a "hop" is a close adjacency where existing capabilities and market position still give you a right to win, while a "leap" is a distant bet, usually into a new technology or business model, that's justified only when the core business faces a structural threat. Most growth should come from hops because leaps that go too far, too early, or too late are all costly failure modes; reserve leaps for genuine paradigm shifts.

Trade-offs & Nuance

Two very different CEOs taught opposite lessons about internal "rebel" teams

Anagnost describes reporting to two prior Autodesk CEOs with opposite philosophies on internal innovation teams. Carol Bartz allowed only a small number of rebel groups, carefully selecting which people could be trusted to push boundaries, and paired that with strong commercial discipline and accountability for actually reaching customers. Carl Bass took the opposite approach, letting "a thousand flowers bloom" with far more experimental teams and rarely pruning any of them, which stressed the company's ability to execute even as it encouraged innovation. Anagnost says his own leadership blends both lessons: Autodesk now scales new businesses more readily because of a cultural willingness to kill its own products, a discipline he credits to learning from both extremes rather than adopting either one wholesale.

Common Mistakes

Mistake: Treating impulsiveness as boldness

Anagnost distinguishes sharply between impulsiveness (acting without sufficient thought, appropriate only in genuine split-second, high-stakes situations) and decisiveness (acting boldly after doing enough "scholarship" to understand the situation). He says he regularly has people try to convince him that impulsiveness is a virtue, conflating it with boldness, and warns that treating a lucky impulsive decision as a repeatable strategy is a mistake; you got lucky once, but you don't want to "play luck on big long" decisions.

Practical Application

Run every kill-or-keep decision through the same three questions

Before funding or continuing any product, business line, or initiative, ask: do we have a genuine right to win here (something specific to our position or capabilities), how adjacent is this to what we already do well, and do we actually have the team and product vision to execute it. Applying this consistently, rather than deciding case by case on gut feel, is what let Anagnost make dozens of kill decisions quickly and be right roughly 70% of the time by his own estimate.

Fix scaling infrastructure before your next big product push

• Do: audit whether your billing, delivery, or operational infrastructure can actually support getting a great product to both existing and new customers at the scale you're planning for.
• Then: prioritize fixing that infrastructure even when it's unglamorous and invisible to customers, before investing further in new product development on top of it.
• Why it works: Anagnost's experience was that no amount of product quality mattered if the underlying system couldn't get that product to customers or scale to new ones.

Pair every hard decision with a repeated explanation of why

When making an unpopular but necessary call, especially one affecting people's jobs or pet projects, over-communicate the reasoning behind it repeatedly rather than announcing it once and moving on. Leave room for people to challenge and shape the specific implementation details, but hold firm and consistent on the underlying "why," since Anagnost credits this repetition with rebuilding trust across a company split roughly in half after a contentious leadership transition.

Questions to Consider

• If we applied a right-to-win, adjacency, and team-capability test to every current product or initiative, which ones would fail all three questions, and are we still funding them anyway out of habit or attachment?
• Are we about to take a "leap" into a distant new market or technology when the situation actually only calls for a "hop" into something closer to what we already do well?
• When we last made an unpopular but necessary decision (a cancellation, a reorg, a strategy reversal), did we communicate the why behind it repeatedly and specifically, or did we announce it once and expect trust to follow automatically?
• Is there a piece of unglamorous internal infrastructure (billing, onboarding, support tooling) that is quietly capping how much value our best product work can actually reach customers with?

Bottom Line

Andrew Anagnost's approach to leading Autodesk through a major pivot rests on three disciplines: a simple, repeatable test for deciding what to kill, a hop-versus-leap framework for how far to stretch into new markets, and a leadership model where courage and empathy work together rather than in opposition. None of these are complicated ideas, but applying them consistently, rather than deciding case by case on instinct, is what let him rebuild both the company's infrastructure and its internal trust after a bruising leadership transition.

Notable Quotes

"It's expensive to use a horizontal tool to solve vertical problems." (Andrew Anagnost)

"Courage without empathy is just meanness and nastiness... empathy is the way you glue courage with an outcome." (Andrew Anagnost)

"Impulsiveness is not a virtue." (Andrew Anagnost)