Dissolve the Roles: Build Small Builder Teams
Shrink teams and drop role labels so AI-empowered individuals own the whole problem.
- Difficulty
- Advanced
- Time to result
- ~months to results
- Steps
- 4
- Confidence
- 90%
Zhuo's counterintuitive practice is to deliberately leave traditional roles unfilled — she does not hire product managers at Sundial — so that whoever is on the team is forced to own product definition, communication, and user value themselves. Because AI lets one person do many jobs at a 60-70th-percentile level, small 'builder' teams without rigid lanes produce faster, better, happier work. The constraint of a missing role is the mechanism that drives ownership.
Origin
Julie Zhuo's operating practice at her startup Sundial, contrasted with traditional Silicon Valley team composition.
Core principles
- 01AI empowers each individual to do many jobs themselves, dropping role boundaries
- 02A missing role forces the remaining people to take ownership rather than delegate into a lane
- 03Smaller teams empowered to act on their specific context beat large teams following high-level rules
- 04The relevant question is which skills the team needs, not which role titles to fill
- 05Early investment is slower as people learn, but well-roundedness pays off long-term
How to run it
- 1
Ask which skills the work needs, not which roles
List the actual capabilities required (product thinking, front-end, data) instead of defaulting to the standard PM-designer-three-engineers script.
Watch out If the team genuinely lacks a needed skill (e.g., nobody can think through product requirements), you must still supplement with someone who has it.
- 2
Deliberately leave a role unfilled
Keep the team small and omit a traditional role so ownership of that responsibility falls to whoever is there.
Pro tip Without a PM, engineers realize 'communication's up to me' and product definition becomes their charter.
Watch out Do not expect everyone to be good at everything overnight; it is less about the role and more about the context.
- 3
Have people use AI to cover unfamiliar territory
Encourage a builder to tackle the front-end or the analysis they haven't done, using AI to get up to speed, with a specialist reviewing.
Pro tip Get a specialist to review the code or give high-level guidance rather than doing it for them.
Watch out Things take longer at first while people learn; budget for that investment.
- 4
Bring in specialists only for heavy, specialized needs
When a project is genuinely, say, front-end-heavy, still pull in someone more specialized for that particular slice.
In the wild
Zhuo expected to need many PMs but found that when an engineer has a PM to delegate to, they punt product definition. Removing the role made engineers own 'figuring out how we get greatest value to users.'
→ Better, faster work and happier employees who feel empowered to create the thing they want.
Instead of slotting a front-end specialist onto a project with some front-end work, she tells engineers 'you're a builder... you can probably figure that out, use AI to help you.'
→ More well-rounded people who can take on many more pieces on their own over time.
Common mistakes
Defaulting to the standard role script
Automatically deciding you 'need a PM, need a designer, need three engineers' fills lanes people then hide behind, instead of matching actual skills to the specific context.
Assuming everyone must do everything
The model is not that each person masters every skill; if the team truly lacks a required capability you must supplement it, or the outcome suffers.
Is it for you?
Best for
Founders and team leads building lean, AI-native teams who want more ownership per person
Not ideal for
Deeply specialized projects, or teams with no one capable of covering a critical missing skill
From the transcript
“we need to dissolve the boundaries of these traditional roles”
“we can drop all of these different role distinctions and call ourselves builders”
“we thought trai we would need a bunch of product managers. It's turned out that actually um if you don't have a product manager”
“we don't have product manager on the team. Okay, so communication's up to me.”
“if we actually make teams smaller and we kind of even like in the past, you know, prei like just have fewer of these”
From the episode
From managing people to managing AI: The leadership skills everyone needs now
Julie Zhuo (Facebook VP, Sundial CEO, The Making of a Manage