For years, Guilds by FirstMark has brought together some of the best company-builders in the world through private, C-Suite communities across product, engineering, finance, GTM, people, and operations. This year, for the first time, we gathered current and future members under one roof for Guilds Summit: a cross-functional conversation about what it takes to build enduring companies in the AI era.

The premise was simple: no function is transforming alone. Product velocity changes finance. AI-native GTM changes hiring. Agentic development changes management. And the companies that win will not be the ones that adopt AI as a side project – they will be the ones that redesign how work gets done.

In keeping with Chatham House Rules, this recap distills some of the biggest lessons from the day for those who could not attend.

Faster is not the same as better

Across the Product & Engineering track, one warning kept resurfacing: AI makes it easier to ship, but not easier to know what should be shipped.

Jon Hyman, Cofounder and CTO of Braze, called the current moment an “AI reality distortion field.” Agents are everywhere. Omniboxes are everywhere. Product teams can push more code than ever. But as Jon put it, “Just ’cause you can push stuff out doesn’t mean you should.”

That theme carried through the day. Ely Kahn shared how at Okta, the company moved from a months-long, waterfall-style product process to an AI-native product development lifecycle in under 90 days. In one pilot, Okta cut time-to-initial-customer-feedback from 45 days to 11, and delivered a major customer-requested feature in 58 days instead of the typical 120.

At Factory, Eno Reyes argued that code generation is only one node in the software development loop. The real opportunity is to redesign the full “software factory”: signals, planning, coding, review, QA, security, deployment, monitoring, and feedback. When AI touches every step, the hard questions are no longer purely technical. They become questions of measurement, governance, and judgment.

Daytona’s Ivan Burazin made the infrastructure case underneath it all: AI-native workloads do not behave like traditional workloads. They spike differently, consume differently, and need infrastructure built for agents, not just humans.

Li Fan of Circle made a similar argument from the customer side: if agents become economic actors, they will need infrastructure built for machine-to-machine transactions, not human checkout flows.

Across the track, the strongest message was not “ship more.” It was “ship with better judgment.” AI raises the ceiling on velocity. It also raises the cost of building the wrong thing faster.

CFOs are becoming system designers, not just stewards of capital

The Finance track made clear that the modern CFO is no longer just managing budgets. The role is becoming a systems-design function: incentives, operating cadence, capital allocation, and clarity.

Will Petrie of Ramp captured it with a simple phrase: the CFO as “chief clarity officer.” Finance still needs to understand the business seven layers down, but AI changes the speed and surface area of that understanding. Ramp’s finance team is building internal tools, prototyping workflows, and using AI to remove bottlenecks across G&A.

Mary Liu of Runway pushed the philosophy further: budget is the wrong frame. ROI is the frame. The question should not be, “Is this within budget?” It should be, “Is this worth it?” That shift turns finance from a gatekeeper into a strategic thought partner.

At Legora, David Eckstein showed what this looks like inside an AI-native business. The differentiator is not just the model. It’s the workflow, the context, and the domain expertise wrapped around it. “This concept of a forward deployed engineer, forward deployed lawyer, is really essential in this AI economy,” he said.

The best CFOs are not choosing between growth and discipline. They are asking where intelligence, capital, and people compound fastest.

GTM is becoming a product surface

The GTM track challenged the idea that sales and marketing can keep operating on static playbooks while product and engineering become experiment machines.

Becca Lindquist of Clay framed the shift directly: “In any organization, you’re either shipping code or you’re shipping revenue.” Clay treats GTM like product. Plays move through a task board. Releases go out biweekly. Data quality, automation, and GTM engineering become core infrastructure.

At Brex, Garrett Marker showed how AI can turn the instincts of a 10X seller into systems the whole organization can use. Brex’s narrative platform creates a continuously updated account narrative from customer signals, then makes it available across the GTM organization.

At Anthropic, Michael Hartman emphasized that AI-native GTM is still deeply human. The company is using AI to improve prep, onboarding, and customer understanding. The goal isn’t to replace customer work—it’s to free leaders and sellers to spend more time doing it.

Bilt’s Zoe Oz brought the brand lesson: experimentation still matters. “Rent Day” became more than a rewards moment; it became a marketing platform. Roomies became more than content; it became a media strategy. As Zoe put it, “Most of your ideas SHOULD be rejected.”

In GTM, the advantage is no longer one great campaign. It is the ability to turn signal into repeatable plays before the market catches up.

People leaders are redesigning work without losing the human element

The People & Ops track tackled the most emotional question of the day: what happens to people when we introduce agents and the work changes?

Iain Roberts, formerly of Airbnb, argued that the people function needs to operate more like a product organization. His question to his team was always: who is your customer, what value are you creating, and how will you know if you moved the metric.

Katie Burke of Harvey offered one of the clearest management metaphors of the Summit: lead like a conductor. Managers are not disappearing, but weak middle management is. In an agentic organization, leaders need craft fluency, crisp communication, strong scoping judgment, and proximity to the work.

Lorraine Buhannic of Headway and Carolyn Frey of Hungryroot showed how workforce planning is compressing from annual cycles to monthly or weekly recalibration. Entry-level jobs are changing, but not simply disappearing. The administrative work is falling away. The need to build judgment remains.

Alex Buder Shapiro of Clay warned against recreating old operating models with new tools. Her team is experimenting with more customized development, fewer translation layers, and AI-enabled recruiting that screens people in rather than out.

The lesson was not that AI makes work less human. It was the opposite. As more tasks become automated, judgment, trust, taste, creativity, and hard conversations become the center of the function.

The company itself becomes the product

The most striking part of Guilds Summit was how similar the questions became across functions.

Product leaders asked how to avoid building the wrong thing faster. CFOs asked how to allocate capital when teams can do more with less. GTM leaders asked how to turn signal into systems. People leaders asked how to redesign work without losing the human fabric of the company.

Different functions. Same mandate.

Do not bolt AI onto the old org. Redesign the org around what is now possible.

That is what we heard from 26 operators across Tubi, Braze, Daytona, Glean, Factory, Okta, Circle, Datadog, OpenAI, Legora, Mercury, Ramp, Runway, Shopify, Bilt, Anthropic, Synthesia, Clay, Profound, Brex, Airbnb, Harvey, Headway, Hungryroot, Airwallex, and more. Not a single leader claimed to have the final answer, but the best ones shared a pattern: move fast, stay close to the work, build from real customer or employee pain, measure outcomes, and keep changing the system as the system changes.

The takeaway: in the AI era, the company itself becomes the product and every function has to learn how to ship.

Guilds by FirstMark is our private C-Suite community for executives from the world’s leading technology companies. To learn more, join a Guild, or apply for next year’s Guilds Summit, visit https://guilds.firstmark.com.