The Operator will soon be available as a podcast. Listen on Substack, Apple Podcasts, and Spotify.

Setting the Stage

Hollie Wegman has had a front-row seat to some of the most interesting company-building stories of the last decade. As VP of Marketing at Segment, she helped scale a company all the way to the $3.2B acquisition by Twilio. Hollie also worked with Redpoint Ventures as an Entrepreneur in Residence, advising dozens of startup founders. Most recently, Hollie worked with Attio, the AI-native CRM that raised $52M led by GV, as an advisor and COO.

Hollie recently departed Attio to invest in and advise AI startups. We connected through a mutual connection at Peak XV, and I was keen to learn what someone with Hollie’s breadth of experience sees when she assesses how startups should be built today.

This conversation goes deep on what actually changed about the Operator playbook in the AI era.

From Marketing Leader to COO: Hollie’s journey

Hollie started her career in Marketing, then over time moved into more general company operations topics. Her transition from Marketing Leader to COO happened organically:

“I came into Marketing via Analytics, so I’ve always had an analyze-the-business lens. As a leader of any function, you should be thinking about the company and what they’re trying to accomplish: what market forces and opportunities are in play. When you build that way, there’s a natural drift towards running not just your function, but also running the company.”

At Attio, this drift followed a familiar pattern. She joined as an advisor before the Series A round, working closely with a 15-person team on positioning, pricing, packaging, and launch strategy. As the company found product-market fit and the challenges shifted from finding PMF to scaling, her scope expanded into people, talent, legal, and support.

Asked about the COO as a role blueprint and title, Hollie was very clear:

“I don’t think it matters what position you have in a company. If you’re on a rocket ship, it doesn’t matter which seat you’re sitting in. If you’re willing to take on different projects and learn, the title doesn’t matter at all.”

She frames the business itself in three pillars: engineering/product/design, go-to-market or GTM (marketing, sales, support), and G&A (finance, legal, people). Entering through GTM gave her visibility into the revenue side while bumping into finance, legal, and other operational functions. It’s a natural springboard into broader leadership though Hollie admits that it doesn’t teach you engineering or product.

Running an AI-native startup: Jazz, Not Symphony

One part of operating startups I find most interesting are inner company operations, basically the structure, systems, and rituals that keep a company running. Traditionally, those items would fall into the realm of a COO or Chief of Staff - yet more and more startups have changed the way they operate entirely (check out how Linear thinks about this).

When I asked Hollie to walk me through the cadences, OKRs, and operating structures at Attio, she pushed back on the premise entirely. This was one of the more refreshing parts of our conversation:

“In my earlier career, we were trained that everything must be a rational system. You need OKRs, you need to cascade them, you need notes and actions. All this stuff is really great when you want to orchestrate a big company. But in this new era where AI is automating a lot of tasks, instead of constructing these tightly wound symphonies, you have to think of it like a jazz artist.”

No rigid weekly cadences or cascaded OKR trees. Instead, Hollie describes an environment of radical ownership where people have deep context, carry big ideas autonomously, and are supported to deliver without being told how to do every single thing:

“I like to work in creative teams, solving problems through brainstorms, spending time together synchronously and asynchronously: stand-ups, huddles, one-on-ones but less formulaic. Creating an environment where somebody feels they have a lot of context and they can carry big ideas through.”

On KPIs and particularly North Star metrics, Hollie doesn’t dismiss them but challenges the conventional approach:

“If everything in the organization is orchestrated around a number - say 10 million - well, what if you could have done 20 million? You orient towards something and that’s the most you’re going to hit. You do need North Star metrics, but you have to keep pressure testing what’s possible.”

Even VCs, she says, are on board with this looser approach: they expect to see numbers and know what’s happened, but they’re not asking for multi-year plans. Planning a quarter ahead feels reasonable. A year ahead feels like folly.

The Hardest Part of the COO Role That Nobody Talks About

One of the sharpest moments of our conversation came when I asked Hollie about the unspoken challenges of being a COO:

She describes a psychological tension that any Operator will recognize: the better you do your job, the less visible your contribution becomes. And when people view operational work as table stakes, it can be tough to sustain.

Her advice for avoiding the trap? Build thinking time into your calendar:

“Once a month or so, take a couple of days where you don’t have a day full of meetings. Allow yourself to sit and think. Jump on a huddle with someone, maybe not even in your department. Apply that kind of radical curiosity and reach across the organization to figure out what people are doing. Then, brainstorm, leave space for ideas, challenge the way we do things.”

The engine room metaphor came up naturally in our conversation, and it stuck: a great Operator knows the gears of a company intimately, but they can’t afford to stay inside them all the time.

The AI-Native Org Chart or Hiring Frontier People

Hollie’s perspective on how AI is reshaping the startup org chart comes from operating inside it at Attio and now supporting founders across a portfolio of AI companies.

People, she insists, are still critical yet the ideal hire persona has fundamentally changed:

“The kinds of people we’re looking for are people who almost have a founder-ish perspective. They’re very happy to grab a problem, use AI to build workflows, and amplify their ability to get things done. [They’re] able to jump in, grab other people, create teams to tackle problems [...], crossing departmental boundaries.”

She distills the ideal profile into three attributes: relentless problem-solving, the ability to create their own operating principles on the fly, and bold, unexpected ideas paired with the ability to MVP them.

On the GTM side, she sees the biggest shifts in how much a small team can now accomplish. Writing, workflows, website building, customer support are all areas where AI has created what she calls “a huge assistance layer.” At Attio, a large percentage of support volume could be handled by AI with the right operating practices in place.

Hollie captured this shift in a LinkedIn post “traditional work is 9-to-5, AI startups work 9/96, agents work 24/7”, hinting at a deeper structural shift. Software is no longer built by humans for humans. Increasingly, it’s built for agents to use. And that changes everything about how you think about product, distribution, and operations.

“We’re All Operators Now”: The Changing Role of the COO

Having worked across company stages and verticals, I was keen to better understand the changing COO role. Here, Hollie shared perhaps the most provocative take from our conversation:

“I think the role of every person working in business is changing. This is a hot take, but they’re becoming more COO-like. As AI takes a lot of automation out of the work, the rigid team structures lighten. We all have to work cross-functionally. We all have to work for the company’s goals rather than necessarily in Finance, or in Marketing.”

The traditional core responsibilities of a COO, staying above the day-to-day, understanding macro trends, figuring out how to move the whole business, are becoming a baseline expectation for everyone. AI gives everyone the tools to operate that way. The question is whether people have the right mindset and agency.

Lessons from Segment, Miro, and Hex

Hollie has worked with and advised companies at wildly different stages and in different categories. At Segment, she was there right up to the Twilio acquisition. At Hex, she got an early look at how AI could be woven into the product experience. At Around (later acquired by Miro), she worked closely with a small team. Across all of them, she sees two common threads in the ones that succeed: deep customer focus, and the ability to react to the market fast:

“All of the people in leadership in these companies are really well tuned to the market and really well tuned to the customer. And they’re all brand-centric: they want to create a feeling about why it’s exciting to use the product, not just a rational argument.”

At Segment specifically, the quality of the team left a lasting impression. She describes it as being surrounded by people so talented that simply listening and spending time with them put you on a constant learning journey. At Hex, the takeaway was slightly different: seeing early how AI could become native to the product experience rather than an afterthought. Each company taught her something distinct, and she carries those lessons across every advisory and investment she makes today.

Given Hollie’s fantastic marketing background, I wanted to understand how she sees the role of a strong brand in the software space. Hollie offered a compelling duality:

“As long as a human is buying software, there’s always not just a rational frame. Brand matters because brand is what feeling you have about a company.
There is a second world: if agents are the primary users of software, are agents eventually going to become the primary consumers of software? In which case, I don’t think an agent is going to care about the brand. It’s going to ruthlessly pick on who has the best price and gets the job done.”

In the near term, however, and today’s buying patterns, Hollie is certain that brands are definitely going to matter.

Patterns Across AI Startups

After four years in the COO seat, Hollie has shifted to investing in and advising AI startups. The vantage point is different and in some ways, richer. Instead of being deep in the gears of one company, she’s now looking across a portfolio and spotting patterns that are hard to see from the inside.

Her thesis is focused on the infrastructure layer: AI governance, AI search, and companies rethinking what software even looks like from first principles:

Across the companies she works with, she’s found that what separates the great Operators from the okay ones has nothing to do with which operating framework they use. Some run tight OKR processes, others rely on pure improvisation. The difference is something harder to pattern-match for: intuition about when to switch modes.

“The pattern is there isn’t a pattern. The ones who [succeed] have the intuition to know when to switch modes from very structured to very unstructured and go after it relentlessly. Strong opinions, loosely held.”

She describes a tension in advisory work that any Operator who has moved between building and advising will recognize: you get to stay on the innovation layer without getting sucked into the weeds, but you know the details matter:

“There’s this temptation when you’re in advisory. You think, ‘if I could just get in the gears more, I could do more for the company.’ But in this era, it’s actually nicer to stay out of the operations completely.”

Her reasoning is pragmatic: the mechanistic, repetitive operational work is increasingly becoming commoditized as AI absorbs it. The strategic layer above, reading the market, connecting dots across companies, challenging assumptions from the outside, is where the real leverage sits right now.

Advice for Fellow Operators

Since The Operator newsletter is aimed at a technically-savvy Generalist audience, I was naturally curious to ask for Hollie’s advice for fellow Operators. Interestingly, Hollie writes her own newsletter called “Dear Operator,” and her definition of the audience tells you everything about how she sees the future:

“Dear Operator isn’t necessarily aimed at a COO or somebody in operations. The way I see the future is that everybody who works in a company is an Operator of AI. I’m thinking about the people working alongside AI for now, and eventually the people who are going to be the human in the loop for the agents.”

Her advice for staying relevant? It’s less about specific skills and more about a disposition:

“Curiosity. ‘How does that work? If I got that tool, how would it work? What if I got this other tool?’ It’s this deep curiosity and fearlessness. We can build software ourselves without being an engineer now. If you’re curious and fearless, those characteristics will carry you through.”

On career progression specifically, she has one clear message:

“I think people should stop thinking about that. I’ve taken roles called marketing that weren’t marketing. Think about how to build companies, think like a founder, and I think you’ll go very far.”

Thank You

Thanks for reading this issue of The Operator.

Hollie’s career arc, from marketing to working alongside VCs at Redpoint to COO of one of the fastest-growing AI startups to advisor/investor is a masterclass in the kind of adaptability that is so required in the AI-age. I especially appreciated her level of abstraction on where AI startups are going and how to stay relevant.

Big thanks to Hollie for this fantastic conversation. You can follow her work on Substack and connect with her on LinkedIn.

See you soon,

Jan

Setting the Stage

Scale AI is the dominant player in human data labeling for artificial intelligence. The company powers the training of frontier AI models (think OpenAI, Anthropic, Meta) by providing the human feedback and data annotation that make LLMs work.

Founded by Alexandr Wang and Lucy Guo in 2016, Scale AI has grown into one of AI’s most critical infrastructure companies. In 2025, Meta bought a 49% stake in Scale AI for $14B and integrated some of their key technologies and teams.

For this conversation, I sat with Adam Gramling, who joined Scale AI in 2022 and spent three and a half years running some of the company’s largest enterprise accounts. Adam initially joined as an Engagement Manager and eventually rose to start and lead both Technical Account Management and Program Management for GenAI. Today, Adam runs SR² Ventures, his own AI-focused investment fund and advisory practice.

Adam’s background spans management consulting, corporate strategy, and a Chief of Staff role in a real estate development company. Today, his portfolio at SR² includes investments in YC companies including several robotics and aerospace startups.

As a fellow generalist navigating the AI space, I found Adam’s perspective on client delivery, onboarding frameworks, and pattern-matching from Operator to Investor invaluable. I hope you enjoy this issue as much as I did.

From Engineering to Business

Adam’s path to land at Scale AI wasn’t really linear. Following the advice of some Silicon Valley executives he had interviewed, Adam studied industrial and systems engineering at USC. Post-graduation, he went into Management Consulting to work across data analytics projects. When an unexpected opportunity came up, Adam became Chief of Staff at a leading real estate development company:

“I moved out to New York, wore a suit and tie, and worked on Wall Street. That was a very interesting role, quite different from working on data projects at tech companies in San Francisco. It was fast paced and I had to get up to speed, research key initiatives, and deliver excellent work on very tight timelines.”

When the macro environment shifted, he moved to NetApp, a data infrastructure company, to work in corporate strategy. Adam deepened his understanding of data and AI. He wasn’t looking to leave when Scale AI reached out.

Joining Scale AI: Data as the backbone of the AI-age

Interviewing with Scale AI, Adam wanted to ensure to find a role that would match his previous background and ambitions. He ended up joining as an Engagement Manager for Scale’s largest customer:

“I absolutely loved that role. You were essentially a mini-CEO running everything for that project: your revenue, cost, dealing with client management, building out new product features. I had a lot of fun doing that.”

What Adam found at Scale AI was a company moving at breakneck speed in a market that was about to explode. When ChatGPT went viral in November 2022, the demand for training data skyrocketed. Scale AI was poised for growth and needed to succeed quickly in a competitive environment.

Speed as a Core Value

Asked for his biggest learnings from Scale AI, Adam’s answer was immediate: speed.

The second lesson was about transparency:

“Be very crisp. Know your numbers, know your data, know why things are working. And don’t try to BS people, because if you do, they will tear you apart.”

The Four-Week Onboarding Framework

As Scale AI was in hyper-growth mode, Adam and his peers needed to onboard new team members at record speed. Working with existing customers and successfully onboarding new ones required a steady flow of new talent.

One of the most tactical insights Adam shared was his approach to ramping up new hires. Early on at Scale AI, the company threw people into complex customer relationships without context and paid for it with high turnover and frustrated employees.

Adam helped develop a structured four-week ramp-up plan:

• Week 1: Don’t do any work. Just learn about the company and meet people.

• Week 2: Start learning about your specific role (think, Account Management, Sales).

• Week 3: Learn about your specific account.

• Week 4: Begin transitioning to autonomy.

Client Delivery as a Services Business

Much has been written recently about client delivery in AI-first companies. A16Z coined it the “The Palantirization of everything”, referring to the Forward-Deployed Engineering model of engineers working closely with customers in project-based settings to provide value and customize existing products. Generally speaking, a core group of engineers and Engagement Leaders or Account Managers would ensure that customers can fully derive value from a product or service.

Working with some of the biggest enterprises in the world, Scale AI followed a similar approach and built up close relationships across their customers. Having managed enterprise relationships at one of AI’s most prominent infrastructure companies, Adam had strong opinions on what separates good Operators from great ones in client-facing roles:

“You are in a services business. Think about going to a nice restaurant: they greet you, they try to make it a welcoming experience. That’s how you need to think about customers. You are representing the company, and you want them to have a fantastic experience.”

Adam also emphasized the importance of being physically present:

“Be on-site with your customers at least once a quarter, ideally twice: once for a mid-quarter visit and once for the QBR. It makes such a difference. I see a lot of younger individuals opting for Zoom calls and my advice has always been the same: hop on a plane and get face time.”

The Future of Human Data

The human data labeling industry, which powers the training of every major AI model, has shifted dramatically since GPT-3.5 went viral in November 2022, Adam remembers:

“Back then, any data you could get your hands on, someone would buy. Now, we’ve moved towards significantly more specialized datasets: advanced math and physics problems, robotics, RL environments, just more complex subjects.”

Adam predicts significant consolidation in the next 12-24 months:

“The barrier to entry historically has been so low. As long as you have a customer, you can find someone to label data. This business has been around for decades. A lot of people saw Alexander Wang be successful and wanted to do the same thing.”

Adam actually sees a much larger opportunity for data labeling in the physical and robotics world. He’s leading with a major question: will companies like Figure and 1X handle data labeling themselves, or go to outside vendors? This and related questions motivated Adam to return to his Engineering roots for his new venture.

Shifting from Operator to Investor

After Scale’s acquisition, Adam decided to formalize his angel investing activities he had been doing for nearly two years. He launched SR² Ventures, and has since invested in 15-20 companies. His investment focus is on startups leveraging human data and robotics while he occasionally helps with hands-on GTM-sparring.

His selection criteria are simple:

“Everyone needs to be a high ARR-per-head business. You need to be a lean organization. If you just throw people at the problem, you won't be able to compete.”

For AI companies specifically, he looks for high-quality, specialized data plays with diversified product offerings:

“Being an AI-wrapper is not enough. You need a unique advantage, for example a competitive moat and paying customers. I see many people building features that do not stand out in the crowd. Maybe you’ll make money for a few months, but it's a matter of time before someone overtakes you.”

Primed by his client-facing work at Scale AI, Adam sees the biggest mistake of early-stage founding teams in building without talking to customers:

“Many companies I speak with base their product decisions on a couple, select conversations. Especially with AI, people get excited about new products easily. But fundamentally, you still need to solve critical problems for the customer. A nice-to-have is very different from solving real business problems, and the willingness to pay long term will differ vastly.”

Leveling up

Asked for his favorite resources and sounding boards, Adam didn’t hesitate and shared the following:

1. How to Win Friends and Influence People, an all-time classic.

2. Good to Great for understanding what separates good companies from exceptional ones, a framework he still returns to periodically.

3. Ben’s Bites, a newsletter for staying up to date in AI.

Thank You

Thanks for reading this issue of The Operator.

I enjoyed learning about Adam’s path from Chief of Staff to Scale AI to running his own angel fund. His framing of client delivery as a “services business” and his structured onboarding approach are frameworks any operator can apply immediately.

Big thanks to Adam Gramling for taking the time to share his journey!

By the way: I’ve just launched my subscriber chat. Would love to see some of you there for discussions and wishes/recommendations for next topics.

See you soon,

Jan

Setting the Stage

Runway is one of the leading Gen AI startups in the media space, developing models and tools for professional content production. In just a few years, Runway has become one of the companies in the Gen AI space working with some of the most prestigious Hollywood film studios.

Most recently, Runway raised a $308M Series D round at a $3B valuation led by General Atlantic, and is estimated to have reached a $300M revenue run rate by the end of 2025. Founded by Cristóbal Valenzuela, Alejandro Matamala, and Anastasis Germanidis who met at New York University, Runway enables film studios to reduce production costs and develop movies faster.

For this conversation, I sat with Michelle Kwon who joined Runway in early 2021, when it was just a six-person team. Over the past 5 years, Michelle helped turn Runway into one of the leading research companies in the Gen AI video space. Today, Michelle leads Operations and Partnerships at Runway as the company scales towards 150 people.

Michelle is also an active angel investor and, together with Lenny and Dan, runs the AirAngels syndicate and fund. She shares how angel investing has sharpened her perspective as an Operator.

As a fellow Airbnb alum, it has been great chatting with Michelle and learning from her journey building a company with proprietary AI-models. I hope you enjoy this issue as much as I did.

From Hedge Funds to Hollywood: Michelle's Path to Runway

Michelle started at a hedge fund in 2014, working with a portfolio manager making early investments in fintech unicorns and marketplaces. She was excited by the pace and trajectory of emerging tech companies.

When the opportunity came, Michelle applied online and landed a role at Airbnb as they were just preparing to roll out the Trips (today, Experiences) product:

“I felt in 2016, Airbnb was the company and the place to be. One of the things I really enjoyed was that Brian Chesky was so visible. He was constantly working in the area that the Trips team was sitting in. He was there on the weekends, with us, too.”

Michelle joined in Business Affairs, which was a cross-functional team that helped scale the cities and markets that Airbnb was going to launch in. She worked from pre-launch to successful expansion in multiple regions.

A few years later, she left Airbnb for business school and started working on the AirAngels syndicate, which turned into a formalized fund in 2020. Around the same time, Michelle got to know about Runway:

“I got introduced to Runway through one of our early investors, who said that I wasn’t really a fit, but recommended to chat with the founders nevertheless.”

Back in 2020, Runway was a small team of six people. Gen AI was in its infancy with little buzz to it. Still, Michelle saw something in the founding team, moved back to New York, and joined in the beginning of 2021 to set up all non-technical functions.

Joining Runway: Looking for Outliers

With the benefit of hindsight and today’s success, joining Runway seems like a no-brainer. However, when Michelle joined, the trajectory of the company was far less certain. Asked for a decision-framework, Michelle credits the mental model from Alfred Lin of Sequoia that she heard through the AirAngels speaker series:

“[Alfred] said, when they’re looking at a company to invest in, there’s product, team, and market. If each of those three things are good, investors will get, at best, average returns. One of those three things has to be an outlier, has to be something that is not prototypical or seems a little bit off. And that’s where you get real outlier returns.”

When Michelle interviewed with Runway, the outlier was clear: the founding team.

“The founders were all immigrants. They had no background in Silicon Valley or in tech, which was very similar to how it was with the Airbnb co-founders. When I was interviewing, I understood maybe 50% of what they were telling me. There was just so much terminology that I wasn’t familiar with.”

The technical depth was important to Michelle as she was looking for a company with unfair industry experience and a defensible moat. Even if the company did not work out, she’d get an incredible learning experience.

Life at Runway

Michelle joined as Head of Operations and Finance to build up all non-technical functions. She quickly realized she could only focus on one main priority at a time, so she changed from one critical business area to another. One of the first projects she tackled was hiring:

“When I first joined, we were desperately trying to hire engineers, and no engineer would answer our DMs or LinkedIn outreach. So I was our first recruiter and first Head of People.”

Michelle spent a large part of the first year on hiring while building foundational operations: from payroll to the first office, all the way to managing investor relations. Once the team realized hiring wasn’t moving fast enough, they brought in their first Head of People. Together, along with the three co-founders and Michelle, they formed the initial executive leadership team.

Working on the most pressing projects, then hiring for that function, has been a main theme of Michelle’s time at Runway:

The same dynamic held true after the team raised a Series B in 2022 and started focusing much more on Marketing and PR. When Runway released their first paying product in early 2023, Michelle acted as the first salesperson before hiring a commercial team. After the successful Series C fundraise later in the same year, Michelle helped find the right candidate to bring the Finance and Legal functions to the next stage.

The Yearly Offsite

One of Michelle’s key learnings has been the importance of regularly stepping back to reassess how the company operates:

“We have a yearly offsite where we revisit how we run the company and how we’re organized. We’re still small, only around 150 people, but we have people who have just joined us, and we recently went through one of those reorgs. That’s a very scary term for a lot of people.”

Michelle, however, refers to reorgs in the best sense: staying nimble as the company grows. What worked at 10 people doesn’t work at 100 people. Michelle describes Runway as having “strong convictions, loosely held”. The team will test something, form an opinion, and if it doesn’t work, change it and try again:

“In the age of AI, speed is everything, so if the organization is not oriented in a way that allows you to run very quickly, make decisions, ship, test, do all the prototypical things, you won’t be successful.”

Adjusting Culture to the Business Reality

Coming from Airbnb where core values were deeply embedded, Michelle knew first-hand the importance of operating under the same assumptions. The Runway co-founders started out with eight core values, which ended up being excessive for a 20-person team. Later, they cut it to four, then evolved into a two-part framework:

1. Smart with Heart: Hire capable people who are genuinely good. No egos, no politics. People constantly chip in and help each other out.

2. Bias Towards Action: If you have an idea, just do it.

Michelle loves how Jensen Huang frames this: “He doesn’t wear a watch because the best time is now. We are already behind. We have to ship. We have to build.”

Runway’s recent success suggests the framework is working well in providing focus and clarity.

Partnerships as the GTM Engine

As Head of Operations and Partnerships, Michelle sits at a unique intersection. While many AI companies focus purely on direct sales, self-serve product-led growth or monetization via API, Runway has built a partnership-first approach that aligns well with their overall business strategy.

Michelle breaks partnerships into two categories:

1. First of Its Kind: Groundbreaking partnerships that create entirely new models. The Lionsgate partnership was the defining example as it was the first major Hollywood studio to partner with an AI company. The partnership combines a customer relationship with data licensing: Runway built a custom model trained on Lionsgate’s +20,000 title catalog (John Wick, Hunger Games), which Lionsgate filmmakers can then use.
   
   Another first of its kind was the partnership with Canva that Runway published in 2023. APIs in the Gen AI space were not widely distributed yet, so Runway ensured the full availability of their video models to Canva customers. This also helped Runway develop their global API which is now the default.

2. Customer Plus: These are large enterprise customers where Runway goes well-beyond providing access to their API. Similar to Palantir and many other AI-companies, Runway uses a team called ‘forward-deployed creative’ to spend significant time helping customers understand the product, create assets, or specialize tooling. While many of those relationships are confidential, Netflix is known to have used Runway’s models in some of their recent productions.
   

Runway’s partnership-focused GTM-motion seems to work well as the company gains customers across verticals well beyond entertainment.

The Future: World Models and Real-Time Generation

Asked for Runway’s outlook for 2026 and major research breakthroughs in the Gen AI space, Michelle calls out two areas:

1. General World Models: Runway announced this initiative two years ago: building AI models that can simulate real-world scenarios. With Runway being the first company to make video models commercially available, the team has invested into being at the frontier of such models. Michelle expects a major jump in quality in the coming two years.

2. Real-Time Video Generation: When Runway launched their Gen-1 model in 2023, it took 3-4 minutes to generate a 4-second video. Today, it takes roughly 10 seconds to create a 15-second video. In the next 12 months, Michelle expects this to decrease to milliseconds with a large associated impact for Runway and its customers:

“Then you’re interacting with and are able to generate video as you’re going along. That really changes the way you can interact with the media.”

Michelle is confident that entirely new use cases would be unlocked - the same way that video generation models have reached the gaming space, for example.

To execute on that strategy, Runway is looking to extend their research- and engineering team, while continuing to make key GTM hires.

The Operator-Investor Flywheel

Michelle’s angel investing with AirAngels provides an interesting counterpoint to her operating role. With dozens of early-stage investments under her belt, she has seen hundreds of pitches.

On whether investing makes her a better operator:

“It does give me a lot of fire sometimes because I see that there are 100 companies that are building towards the same thing that we are trying to build, so we’ve got to move quickly.”

At the same time, Michelle shares that being a hands-on Operator has made her much more discerning in finding conviction for an investment:

What she focuses on above all is founder obsession:

“They don’t have to be experts per se in what they’re doing, but I do think that an obsession with what you’re building is so important. You have to really be an outlier as a founder because the dedication it takes to run a business through highs and lows is a lot.”

With more Operators becoming angel investors and dedicated programs like First Round’s Angel Track to lend support and guidance, I found Michelle’s account very helpful.

Getting Feedback

To connect with fellow Operators, Michelle has been considering joining Molly Graham’s Glue Club. To date, however, Michelle’s best sounding board has been her partner:

“To be honest, my husband is my go-to partner. He does not work in startups but in Finance, so he analyzes businesses all the time. Companies that have gone public get a lot of scrutiny from investors, and there are some best-in-class examples. I’m not spending my time looking into that - that’s what my husband does. So he knows what great companies look like.”

That outside perspective less focused on the day-to-day operations but more on the bigger picture has been invaluable to Michelle as she has scaled herself with Runway.

Thank You

Thanks for reading this issue of The Operator.

I thoroughly enjoyed learning about Michelle’s path from Airbnb to Runway and her experience being an active angel investor. If Michelle’s journey resonated with you or sparked ideas about scaling operations in AI companies, I’d love to hear from you.

Big thanks to Michelle for taking the time to share her journey building Runway!

See you soon,

Jan

The Journey

TrueLayer is one of Europe’s leading Pay by Bank networks, powering payments infrastructure for companies like Revolut, Just Eat Takeaway, and Ryanair. Founded in London in 2016, TrueLayer has raised over $320M across multiple funding rounds and processes billions in payments.

The company’s growth trajectory, however, hasn’t always been linear. When the macro environment shifted in 2022, TrueLayer needed to cut costs - a story familiar to many operators navigating the shift from growth at all costs to capital efficiency.

For this conversation, I had a chance to speak with Jiun Yi Tan. She joined TrueLayer as a founding sales hire in 2018 and transitioned to Chief of Staff to the CEO during the company’s most challenging period. Today, as Director of Strategic Projects, Jiun Yi continues to tackle TrueLayer’s most complex initiatives.

Jiun Yi shares her account of being a Strategic Operator who has been through the ups and downs of building a resilient business.

Leveraging industry experience to land a job at TrueLayer

Jiun Yi’s journey to join TrueLayer in its early days was based on industry experience and curiosity: at Accenture, her team consulted clients on open banking and had nearly launched their own [open banking] startup with a proprietary open banking product. She followed TrueLayer as the most mature solution in the market and, when they opened a commercial role, took the leap.

As one of the first salespeople on the team, Jiun Yi helped lay the foundation for TrueLayer’s early commercial success:

“One of the biggest moments was signing our Revolut partnership. They were our first European launch partner, and I was really lucky that the team let me work on that deal. Revolut is one of our most successful clients that legitimizes everything we do today. We learned so much from partnering with them.”

Jiun Yi gained invaluable operational and commercial experience building out TrueLayer’s sales playbook. Crucially, however, when the company opened their first Chief of Staff to CEO role, her selection was rooted in deep organizational context, existing trusted relationships, and her established reputation. This foundation made her the ideal internal candidate to empower the CEO, enable the leadership team, and drive organizational scale, a mandate that would soon be tested by a critical market situation.

Becoming Chief of Staff during a market shift

In early 2022, Jiun Yi transitioned into the Chief of Staff role. TrueLayer was scaling rapidly across countries, and during the pandemic the team size had increased considerably. VCs invested generously in hot verticals such as open banking (let’s not talk about crypto or NFTs…) and expected rapid scale, all while caring less about the bottom line.

When the macro environment shifted and the ZIRP era came to an end, TrueLayer needed to shift gears quickly, as Jiun Yi recalls:

“One of the first things I worked on was to operationalize the shift from company focus on payments, volumes and numbers of clients to revenue. We needed to do this in a capital efficient manner, which was brand new at that time. Before then (Series B-C), it was about who can hire or build the fastest, so it gave us a little bit of a whiplash. This context shaped my role for the next two and a half years.”

Borrowing the famous analogy from Ben Horowitz, it was wartime mode, which required Jiun Yi and the rest of the leadership team to take decisive steps:

Jiun Yi worked closely with the executive team to navigate the RIF, committing significant effort to ensure the process was handled with dignity and respect. The team prioritized radical transparency and meticulous communications planning to manage the challenging and delicate cultural shift.

Beyond headcount, the company had to make strategic cuts such as shutting down their Australian entity, and deprioritizing stablecoins and alternative payment projects:

“When you need to rationalize investments and fix your burn rate [fast], some things just can’t be worked on anymore.”

While drastic at the moment, those steps ensured that TrueLayer could focus on its core business while navigating a true market shift.

Rebuilding trust and scaling the Operating System

After adjusting the strategy and team to the new reality of financial prudence, Jiun Yi and the leadership team worked hard to rebuild trust, which required radical transparency:

“We hosted lots of live Q&As where we didn’t prepare the exec team or anything. We just put them on stage and made sure that everyone could ask questions and get answers. As an organization, we’ve always been super transparent about revenue or payment volume, but we also ensured to use the same strategic narrative based on data so that there would be no surprises. I think we’ve managed to build back that trust over the following years.”

With TrueLayer’s strategy and focus fundamentally changing, Jiun Yi was poised to upgrade the company’s operating system to the new environment:

“Whatever we did pre-Series C was good for that period of time, but suddenly the game and the rules had changed. So, we asked ourselves: what is the new Operating System? How do we design our decision-making process? How do we engage the leadership team and give them context to help them be more effective? How do we cascade strategy to the rest of the organization?”

Jiun Yi tackled a key challenge for many ‘Chiefs’ - how information flows inside of a quickly scaling organization, how teams stay aligned and focus on what matters, and how leaders can take clear, informed decisions. She introduced, then refined, several key mechanisms:

1. Company-wide alignment: Monthly all-hands meetings became the drumbeat of strategic communication, with quarterly deep-dives going into more detail.

“We always repeat the metrics that are driving the business, the ones we care about most. Everyone knows what everyone’s working on from an operational perspective, but you [also] want to hear what the leaders are really thinking.”

2. Clear goal-setting:

“We’ve always done annual goal-setting, but I helped introduce and formalize company OKRs as part of goal-setting for measurability. Since it was the first year that revenue became the most important metric for us - we’d moved from caring about transaction volumes to revenue - it was all about elevating that focus, and we wanted the all-hands to refer to that more consistently.”

3. Decision-making clarity: Jiun Yi wanted to ensure leadership followed a clear process in taking decisions. Borrowing from the ‘Amazon memo’, she implemented a three-pager decision process:

“Before that paper is done, no meetings are booked. We didn’t want people to just show up to meetings and share their thoughts. We needed to categorize the kind of feedback required from the exec team for review, and sign-off.”

The strategies Jiun Yi introduced such as OKRs are a key tool of the trade for any Chief as we learned in a previous episode. In the case of TrueLayer, it helped the team stay focused while navigating increased competitiveness.

Scaling the Chief of Staff Function

As the company grew, so did the complexity that Jiun Yi and the leadership had to navigate. Fortunately, TrueLayer ensured that each of their executives would benefit from increased leverage through a dedicated Chief of Staff:

“The CEO, CTO, CCO, and CFO, each one of the execs had their own Chief of Staff. Three of us were internal moves and promotions, and our CFO’s Chief of Staff was hired into the role.”

This structure worked remarkably well:

“It’s really hard to be a one-woman band and try to roll things out across the entire company. I was the conduit between what was decided at the exec level, and my partners in crime ensured their orgs were aligned with the same context.”

Jiun Yi was able to operate in a core group among fellow Chiefs that shared an equally intimate perspective on the business. Through monthly catch-ups and a shared slack channel, they ensured all decisions and communications were aligned well.

The Transition: From Chief of Staff to Director of Strategic Projects

After two and a half years as Chief of Staff, Jiun Yi had tackled several strategic challenges and had trained her muscle to identify and solve some of the company’s most complex challenges. Together with the CEO and COO, she co-created a new role and transitioned to Director of Strategic Projects. Asked for the difference between both roles, she shares a clear distinction:

“The Chief of Staff role was all about the operating system, helping your exec and leadership team to thrive. It’s about managing the entire company, so things are aligned.

The Strategic Projects role is more about setting the strategy and creating the structure for different initiatives at various parts of the growth stage. I identify gaps in the business and set the vision to maximize the probability of successfully delivering this year and next year’s initiatives with the cross-functional team that we have. It’s a lot more embedded within the business.”

The shift came after important feedback from her CEO:

In her new role, Jiun Yi tackles initiatives that don’t fit neatly into existing org structures, like navigating new UK Authorized Push Payment regulations requiring payment platforms to reimburse 50% of consumer fraud losses, or developing TrueLayer’s AI strategy while balancing her CEO’s enthusiasm with operational reality.

“Our CEO is an AI evangelist - he would AI-ify the entire company today if he could. My job is to be a little more balanced. Based on where TrueLayer is right now, I know what our data is like, what our processes are like, and that maturity levels across departments are different.”

TrueLayer at an Inflection Point

The operational excellence that Jiun Yi and the exec team built during the challenging 2022 period has positioned the company for its next chapter. Through clear strategic decisions and trade-offs, TrueLayer has emerged stronger and reached an inflection point.

In October, TrueLayer announced its first acquisition (subject to regulatory approval) to expand into the Nordics and strengthen its market position as the leading Pay by Bank provider in Europe.

The shift from wartime mode to strategic growth demonstrates what disciplined Operators can achieve: build resilience, weather market shifts, and emerge stronger for the next phase of growth.

Advice for Strategic Operators

Reflecting on her seven-year journey at TrueLayer, Jiun Yi offers first-hand advice for fellow Operators considering following the Chief of Staff path:

1. Recognize the risk:

“It’s quite a high-risk role because you give so much of yourself and your time, and you get really dedicated to somebody else’s or the company’s success. The first thing is to make sure you’re genuinely obsessed with the opportunity and what you’ll learn.”

2. Do your due diligence:

“If it’s not an internal move, you need to understand: what is your principal like? What are they good at? What are they not? Because what they’re not good at is what you’ll be working on. You want to make sure that aligns.”

3. Network authentically:

“Networking really is the best way. Founders are quite time-poor, and they want to work with people who have the same passion for the space. If you are naturally interested, you’ll spot the opportunity. My TrueLayer role didn’t pop up on AngelList. I was looking for it because I knew of the company.”

4. Stay energized:

“The main guiding principle I have, and maybe why I’ve managed to stay in a high-growth, high-scale company for this long, is because I’ve always sought the elements that energize me. It’s so important to know what you’re good at and what actually energizes you, because when you don’t have that alignment, it’s a recipe for burnout.”

In terms of resources, Jiun Yi recommends joining the Chief of Staff Network to network and learn from peers, and reading Molly Graham’s blogs for lessons on scaling high growth organizations.

Thank You

Thanks for reading this issue of The Operator.

TrueLayer’s journey from hot open banking contender to hyper-growth company scaling and correction to operational efficiency offers incredibly valuable lessons to fellow Operators.

I appreciate Jiun Yi’s candor in sharing her wartime stories, cultural shifts, and personal evolution, which have made this conversation super compelling. Thank you, Jiun Yi, for joining.

See you soon,

Jan

Linear is one of today’s most notable software startups building high-quality, nuanced products. Starting out as an issue-tracking software, Linear has become a serious Jira challenger working with F500 and some of today’s most innovative companies.

Linear’s approach to company-building contradicts some of the blitzscaling playbooks startups look up to. Instead, Linear features negative lifetime burn, an all-remote experienced team, and incredible success that most recently led to a $82M Series C round at a $1.25B valuation.

Other interviews with the Linear team have covered their approach to product management or itereating on the perfect messaging.

For this conversation, I wanted to focus on the inner workings at Linear that would intrigue any Strategy & Operations leader.

On a personal note, as a fellow Chief of Staff, Zoe and I have been in touch since a few years, and it has been great seeing her excel at Linear.

From Tactical to Strategic: the evolving CoS role

Zoe joined Linear as employee #12 and first business hire. At such an early stage, tactical execution was the focus:

"Every day, my priorities were extremely clear because it was simply: what is the thing that's on fire that I need to get done today?"

Her early focus split between supporting the product team and ensuring leadership priorities were clear. In addition to that, Zoe took care of ad-hoc items such as hiring when no Talent team was in place yet, and operational items such as Finance, Accounting, Billing.

Today, with a team of almost 100 employees working entirely remote, her focus has shifted to providing unique strategic insights:

"Now it's much more about pausing to be strategic. I'm often thinking about when it makes sense to hire a team for some of the things I'm doing, or whether we as a leadership team are making the right trade-offs in our decisions."

Despite the scaled organization, some of Zoe’s core responsibilities have remained constant: all-hands meetings, investor relations, board preparation, and company planning. Overall, however, with more functional leaders at Linear, she has moved from direct ownership to facilitating the right conversations.

Company-Building at Linear

Hiring Approach and Team Philosophy

From the early days, Linear has been intentional about both what they build and how they build. With a fully remote team already pre-pandemic, Linear’s co-founders set out a clear vision for the company culture they wanted to establish:

"From the very early days, we were looking for people who cared about their craft, were low ego, could flex into different areas, and didn't have strong feelings about what they were working on or their titles."

Essentially, Linear was indexing on self-starters who excelled in their field and cared about their craft, with no place for micromanaging:

"We work in really small teams and independently, sometimes project teams of one or two people. You need to be motivated and know how to unblock yourself. We don't expect extremely fast responses from people, so sometimes it'll take a day to hear back. You need to not be stifled and keep making progress."

Linear established work trials early on as part of their interview process, aiming for a high conversion rate from trial to offer. Throughout a work trial, the Linear team and the applicant work on a real-life project to assess mutual fit:

“We get a lot of insight from candidates that way, and it helps kickstart onboarding as well.”

Staying Lean

As Linear has evolved from issue-tracking tool to enterprise-grade project management software, it has stayed incredibly lean. With the latest Series C funding round, the Linear team is about to cross the 100-employee mark. While the overall trend in startup team size is trending downward, Linear champions succeeding with a proportionally small team.

One of the key metrics Linear tracks is revenue per employee.

"You never want that number to drop. It's not a perfect metric, but ideally, it grows or accelerates over time. We don't get less efficient as a business, which is sometimes counterintuitive for VC-backed startups."

According to Zoe, this philosophy creates "healthy pressure" on every hire:

"Our CEO asks great questions about understanding the person's day-to-day: who they work with, what they are going to do, what value they are going to bring. We really understand what skill sets they're going to unlock for our business."

The approach has evolved slightly as they've grown, but the core principle remains: every hire must demonstrate to increase the company's efficiency, not just its headcount.

Running Linear

Focusing on what matters

Linear’s operational philosophy mirrors their hiring approach of having a lightweight process with maximum clarity. The team does not use quarterly OKRs, extensive KPI dashboards, or rigid weekly business reviews. Instead, they rely on bi-annual planning cycles in the fall and spring, focusing on big milestones while staying flexible.

Linear’s key to success is to keep planning transparent while limiting involvement in decision-making to a small group. This approach requires deep trust with the leadership team, but it has ensured that the team remains strongly focused on what matters most.

In addition to large-scale company planning, monthly all-hands meetings provide deeper updates with demos, keeping the team aligned without over-managing the process.

CoS and COO: Two Distinct Roles

In 2023, Linear hired Cristina Cordova as COO to focus on revenue-generating functions. When Cristina joined as COO, Zoe had to navigate the classic Chief of Staff question: what is the overlap between both roles?

"The Chief of Staff role is to keep the leadership team as productive as possible, and that person is needed separately from the operations person, whose goal is to keep the company running efficiently."

In the case of Linear, Cristina focuses on go-to-market and revenue functions, while Zoe handles internal operations such as team operations, leadership priorities, Finance, HR, and compliance. However, as the company continues to scale, these responsibilities may shift:

"Sometimes the roles get conflated in early-stage companies, but they serve very different purposes. It's important to keep them as two different roles if you're able to."

AI at Linear

Linear works with some of the world’s most innovative AI companies such as OpenAI, Scale AI or Perplexity. In order to help them accelerate their software engineering efforts, the Linear team needs to understand those tools really well. The team launched Linear for Agents where AI agents function as real team members.

Linear encourages all team members, regardless of function, to experiment with AI tools. This internal adoption serves two reasons: it helps the team understand their own users better by experiencing firsthand where AI tools excel and where they still fall short. As Zoe explains:

"We're like a mini case study on what's going on in the rest of the industry. The AI coding tools tools are good, not great, still require a lot of manual review and sometimes can't do what you hope them to do. Some people are early adopters and use it side by side with their code every day and are more productive because of it. Some people are more unsure."

Advice for Fellow Operators

Asked for advice for fellow Operators and Chiefs, Zoe focuses on clarity and cutting out unnecessary distractions:

"You need to take a step back and know your top priorities. They're based on what's most important to your founders, which is based on what's most important to the business. That can cut your to-do list in half."

Chiefs are often exposed to some of the most sensitive topics inside an organization, both good and bad. Succeeding in the role for the long-term requires delegation and letting go:

"You can't let yourself get burned out. A big part of the role, if your company is growing, is being comfortable with giving up parts of the role. You're still one person, so as the company becomes more complex, you have to adapt."

Most importantly, since the effectiveness of a Chief of Staff role is heavily dependent on close relationships, with the executive team and beyond, Zoe encourages fellow Chiefs to over index on that:

Thank you

I hope you enjoyed this edition of The Operator.

Linear's approach to product and company-building offers valuable lessons for fellow Operators, and I hope this conversation with Zoe helped surface some of those insights.

Big thanks to Zoe for taking the time to share your Chief of Staff journey and learnings with us!

See you soon,

Jan

Harry initially didn’t plan to become a Chief of Staff. What drove him wasn’t solving one specific problem or working in one industry - it was thinking in terms of systems and improving execution.

After starting in a technical management consulting role, he craved a more dynamic environment. He ended up working at Techstars in Berlin and met the founder of GTMHub (now Quantive).

"I was much more about how I can plug and play — and use my skills to help facilitate company growth."

Harry joined the company and, as one of the first US hires on the team, started out building the Customer Success function as Quantive was driving US-expansion. Through solving customer issues, building operations from the ground up, and launching the US Go-To-Market strategy, Harry transitioned into a Chief of Staff role.

Inside Harry's Chief of Staff Journey: Scaling at Startup Speed

When Harry joined Quantive, the company was a 20-person startup that had just raised their seed round. Over the next five years, he witnessed - and helped drive - the company’s growth past 400 employees all the way until the Series C.

As Chief of Staff, Harry’s role evolved constantly. In the early days, he tackled whatever gaps existed: supporting customer success, building operational processes, helping recruit the first U.S. sales team. There were no playbooks, just problems to solve and opportunities to shape.

"Every six months, it felt like a brand-new job."

As Quantive scaled, so did Harry’s focus. He shifted from operational execution to coordinating leadership initiatives, leading offsites, and building the systems that would sustain growth. Along the way, he learned that the most valuable Chiefs of Staff aren’t just problem solvers - they’re force multipliers for leadership.

Trust, adaptability, and a bias toward action became his guiding principles.

"If you’re the opposite of your CEO in the ways that matter, that’s often a very good thing."

Making OKRs Work: Lessons from the Inside

At Quantive - a company that builds OKR software - Harry was naturally involved in OKRs, a topic that often falls into the realm of a Chief of Staff. Asked for best practices on OKRs, Harry shares that most OKR programs fail because they’re too rigid, too top-down, or simply disconnected from day-to-day work.

Harry’s approach was different. He focused on helping teams co-create goals that mattered — to the business and to them. That meant fewer objectives, more clarity, and creating space for real ownership.

“OKRs don’t work when they feel like homework. They work when people care about what they’re aiming for.”

Success came from building habits around reflection, focus, and learning - not just tracking metrics. The system Harry built prioritized alignment without killing autonomy.

For operators thinking about OKRs, his advice is simple: start small, stay human, and don’t confuse structure with bureaucracy.

What made OKRs work at Quantive:

• Relentless focus: Limit OKRs to 1–2 key areas. Force tradeoffs.

• Leadership buy-in: Founders and management need to be fully supportive of OKRs.

• Designed for humans: Make sure teams want to participate — not just treat it like another checklist.

• Built-in incentives: Celebrate wins tied to OKRs. Help teams see the personal upside in alignment.

• Iterative improvement: OKRs weren’t static. Teams reviewed and refined them each quarter based on learnings.

"OKRs succeed when they are co-created, not dictated."

From Chief of Staff to Founder: OneTwo Growth

After five years at Quantive, Harry realized he’d scaled himself out of the day-to-day. The gaps he once filled were now covered, and he felt energized to help other companies navigate similar growth moments.

That led him to found OneTwo Growth, where Harry works as a fractional Chief of Staff and Strategy Advisor for startups.

He focuses on helping teams bring clarity and structure to strategy execution, without accumulating bureaucracy. His work often includes:

• Coaching Chiefs of Staff to grow from operators into systems-thinkers

• Helping leadership teams stay aligned as complexity increases

• Designing effective frameworks that let strategy run on autopilot

"Startups don’t need more process - they need the right process that supports momentum."

OneTwo Growth is grounded in a systems-thinking approach: combining people, process, and technology to reach outcomes faster — and with less friction.

Advice for Aspiring Chiefs of Staff

Harry works with both current Chiefs and professionals keen to break into this role. He believes the best Chiefs bring a blend of curiosity, adaptability, and bias for action, not necessarily a perfect resume.

Here’s his advice for breaking in and thriving:

• Your nontraditional background is an advantage: Organizations don’t want clones — they want complementary problem solvers.

• Ask great questions in interviews: How does this company define "success" for the CoS role? Are they looking for a more administrative Chief or for an entirely embedded Chief running key projects?

• Don’t try to be everything: Know your strengths. Lean into them, rather than trying to be perfect at every task.

• Think like a builder, not a fixer: Focus on creating systems and capacity - not just solving today’s fire.

• Invest in your network early: Build relationships with other Chiefs of Staff. Peer learning is invaluable.

"The worst thing you can do is say: 'I can do everything.' That doesn’t set you apart — it dilutes your value."

Harry’s journey shows that the CoS role is what you make of it. Done right, it’s not just a stepping stone but has the potential to be a career-defining experience.

Recommended Resources from Harry

As an active member and mentor in the Chief of Staff domain, Harry recommends the following resources to break into the role and level up:

• Ask a Chief of Staff Slack community by Clara Ma - for connection and real-world advice.

• The Art of Explanation by Ros Atkins - a book with fairly tactical advice to distill complex information for executive communication.

• Your Chief of Staff — a newer community worth checking out, founded by Elise Kennedy.

These resources helped Harry sharpen his thinking and can help others at all stages of the CoS journey.

Thanks for Reading!

Hope you enjoyed this edition of The Operator. If Harry’s story sparked ideas or gave you a new lens on the Chief of Staff role, I’d love to hear from you — just hit reply or drop a comment.

Feel free to subscribe to get the next edition of this newsletter straight to your inbox.

And if you know a Chief of Staff or Operator who’d love this, feel free to forward it to them!

Until next time,

Jan