ExpertsFMBuilding a Team at AI Speed | Harvey’s Maggie Landers
The a16z Show · September 27, 2026
5 Selected Lessons:
heuristic · 3 framework · 1 reasoning_pattern · 1
Radical internal transparency as a trust mechanism
Source
The a16z Show — Building a Team at AI Speed | Harvey’s Maggie Landers (September 27, 2026) · at 9:00 ▸ · at 13:54 ▸ · at 33:30 ▸
When this applies
when leadership needs to keep a rapidly growing, mostly-new workforce aligned and connected to strategy
The takeaway
Default to sharing strategic information, reasoning, and admitted mistakes broadly rather than restricting it to a leadership layer, especially when the workforce is largely new and autonomy is required.
Why it works
Sharing real decision-making context (including mistakes and uncertainty) lets employees understand leadership's reasoning rather than just outcomes, which builds trust and enables faster autonomous decision-making without needing to escalate.
One level up
Trust scales organizations faster than control does.
Where it stops holding
This stops working when information is competitively sensitive enough that broad sharing creates real leak risk, or when the culture isn't mature enough to handle candor about mistakes without panic.
Detail
procedure:
- Post board decks for the whole company after board meetings
- Have founders write and share quarterly/ad-hoc memos on market view and what did/didn't work
- Host founder AMAs (even run by interns) for open Q&A
- Openly discuss mistakes and unknowns, not just wins
benchmarks: {"context":"done at a company scaling toward 2,000 employees with 70%+ tenure under one year"}
rationale: Keeps a mostly-new, globally distributed workforce connected to leadership's thinking and sustains the trust needed for high autonomy.
In their words
We post board decks for the whole company to review after board meetings so they can see what's been shared in those board meetings and host AMAs in Slack so that people can ask questions or get clarification.
Letting go of A-student perfectionism to operate at high pace
Source
The a16z Show — Building a Team at AI Speed | Harvey’s Maggie Landers (September 27, 2026) · at 6:42 ▸
When this applies
when previously successful, high-achieving people enter any new high-ambiguity, high-velocity environment
The takeaway
In novel, high-velocity environments, deliberately shed the 'avoid being wrong' instinct in favor of 'generate many attempts and correct quickly,' because the volume of reps matters more than the accuracy of any single one.
Why it works
People trained to optimize for 'getting it right' on known problems freeze or slow down when problems are novel and undefined; explicitly reframing failure as expected and useful unlocks the volume of attempts needed to find what works.
One level up
Different environments reward different cognitive habits — importing a strength from a different context (perfectionism) can become a liability in a new one.
Where it stops holding
This stops working when the domain actually has known best practices that could be imported rather than reinvented, or when the cost of an individual mistake is catastrophic rather than correctable.
Detail
approach: Recognize incoming hires' prior success patterns (being 'A students'), then explicitly coach a shift toward fast failure, correction, and ego-detachment for collective progress.
signals:
- New hires from traditional law or big tech bring 'right answer' habits
- Experimentation requires failing more than succeeding
- Progress-over-perfection culture as the antidote
rationale: Building something unprecedented requires many fast reps, which is incompatible with needing to always be right.
In their words
And B, there's a lot of experimentation. So you're going to fail more than you're necessarily going to get it right. And so I think the emphasis is on both failing fast and correcting experimentation and also kind of putting that ego aside for kind of the greater good.
Simplicity: ship minimal, iterate from feedback
Source
The a16z Show — Building a Team at AI Speed | Harvey’s Maggie Landers (September 27, 2026) · at 29:16 ▸
When this applies
when launching any new product, process, or initiative under uncertainty
The takeaway
Build the minimum viable version to get real feedback, then let that feedback drive the next iteration.
Why it works
Simple initial versions get real-world feedback faster than fully-built ones, and that feedback is more valuable than internal speculation about what's needed.
One level up
Information from the real world is cheaper to obtain through shipping than through internal deliberation.
Where it stops holding
This stops working when the cost of a flawed early version (reputational, safety, compliance) outweighs the value of faster feedback.
Detail
rule: Don't overbuild — put something simple into the world, get feedback, iterate.
threshold: n/a
rationale: Simplicity paired with decisiveness is why Harvey can move at its pace.
In their words
Simplicity, again, I think is one of the reasons why we are able to move fast, right? It goes hand in hand with the decisiveness, but let's not overbuild. Let's put something out into the world. Let's get feedback. Let's go from there.
Carrot-not-stick office policy still produces high attendance
Source
The a16z Show — Building a Team at AI Speed | Harvey’s Maggie Landers (September 27, 2026) · at 40:42 ▸
When this applies
when deciding whether to enforce or incentivize any behavioral policy (attendance, process compliance, etc.)
The takeaway
Make the desired behavior intrinsically rewarding rather than mandated, and monitor outcomes rather than compliance.
Why it works
When the environment itself (collaboration, learning from peers, energy) is valuable enough, people exceed a minimum policy voluntarily, making enforcement unnecessary and avoiding the trust cost of policing.
One level up
Incentive design that relies on genuine value beats enforcement that relies on monitoring.
Where it stops holding
This stops working when the environment isn't actually valuable enough to pull people in voluntarily, in which case a pure carrot approach yields low compliance.
Detail
rule: Set a hybrid minimum (e.g., three days) but don't police attendance; rely on the value of in-person collaboration to exceed it.
threshold: Official policy: 3 days/week; observed actual: many employees come more than that voluntarily.
rationale: Especially valuable when most employees are new and ramping, since in-person proximity accelerates learning from more-tenured peers.
In their words
We have a hybrid three-day a week policy, but I would say people are in typically more than that. It is very much tarot versus stick when it comes to that, that hybrid in office policy.
Screen for mission-fired-up candidates with a direct 'why this industry' question
Source
The a16z Show — Building a Team at AI Speed | Harvey’s Maggie Landers (September 27, 2026) · at 22:09 ▸ · at 40:42 ▸
When this applies
when hiring for a company whose work is domain-specific and emotionally/mission-charged (vs. generic infrastructure work)
The takeaway
When your work is tied to transforming a specific real-world domain, explicitly test for domain-specific passion, not just general excitement about company growth or AI.
Why it works
Working in a specific, high-stakes domain (vs. general-purpose tech) requires sustained motivation tied to that domain; asking directly what excites a candidate about the industry filters for people whose motivation will survive the grind.
One level up
Generic ambition doesn't substitute for domain-specific conviction when the work requires deep, sustained immersion in that domain.
Where it stops holding
This stops working when the role truly is domain-agnostic (e.g., pure infrastructure/model-building roles not touching the customer), where this screen would be less relevant.
Detail
rule: Ask candidates directly what about the specific industry gets them fired up, and use the quality/depth of their answer as a signal.
threshold: n/a
rationale: Differentiates candidates who are excited about AI/growth generically from those with real, sustaining conviction about transforming this specific industry.
In their words
So, what I feel like I'm hearing is the thing you love to hear most from candidates is what about the legal industry gets you so fired up? I think that's a big part of it.