Guest Episode Episode 08

Why I Left ElevenLabs
to Be Employee #1 at an AI Startup

Megan Standing scaled the people and talent function at ElevenLabs from 30 to over 300 in 18 months, then walked away to be the first employee at Solve AI. A decade in talent, the mechanics of holding the bar in hypergrowth, and why the first hire wasn't an engineer.

About This Episode

Megan Standing spent nearly a decade in talent at Palantir, Typeform and ElevenLabs, where she built and scaled the people and talent function from 30 to over 300 people in 18 months, one of the most watched hypergrowth stories in AI. Last year she left to become employee #1 and Chief of Staff at Solve AI. In this episode she joins Ceci and Bea to unpack the pivot, the mechanics of holding a hiring bar through hypergrowth, and what a talent background gives a Chief of Staff that nothing else does.

What We Cover in This Episode

From a Decade in Talent to Chief of Staff

There was no dramatic trigger. Across Palantir, Typeform and especially ElevenLabs, Megan kept being drawn to the messy operational problems that were company-shaped rather than hiring-shaped. When Steve, who she'd worked with for five years at Palantir, started building Solve AI, the offer was to be the first employee and build from zero across everything. What she really liked about her roles to date was the cross-functionality: legal and new markets, equity plans, budgeting with finance, revenue plans with go-to-market, visa sponsorship policies. She doesn't much believe in job titles: go do the thing where you'll have impact.

What Talent Gives You That Other Chiefs of Staff Don't Have

A decade in talent is a decade of watching companies scale from the inside out: where communication breaks, where leadership training is needed, how incentives shape behaviour. It also builds judgment under uncertainty, because interviews are noisy, references are selective, and people perform differently in different environments. Talent is the machine that determines whether a startup succeeds. And people in talent get a uniquely honest read on the company, because candidates and employees tell them things they will never tell the executives.

What She Had to Unlearn

Heroics. People and talent professionals in hypergrowth say yes to everything. With a remit now spanning finance, legal, compliance, security and GTM, Megan had to learn to say no to things that are important but not now, like dropping everything else because a SOC 2 is the P0 that unblocks enterprise contracts.

ElevenLabs From 30 to 300, Without Lowering the Bar

Megan joined ElevenLabs as employee #28, when there was no people function at all. Mati told her they'd be 60 by end of year; they ended at 130. The quality-control playbook as hiring scaled: invest in interviewer quality, structured feedback, scorecards and rubrics; make hiring managers accountable; write hiring theses reviewed by the founders, each with a success mode, a failure mode and a mitigation; and keep founder involvement, Mati and Piotr interviewed every candidate at final stage. The counterintuitive part is being comfortable slowing down: one exceptional hire outperforms four mediocre ones, and one bad hire compounds. Holding the bar without becoming a cutthroat culture comes down to being explicit about expectations before someone even joins, giving feedback in the moment rather than in six-month cycles, and giving people a genuine, tracked chance to improve.

Recruiting Before and After a Company Gets Famous

Early on it's persuasion: talent mapping, creative outreach, convincing people to take a risk on an unknown Series A company. Once the company is known, inbound explodes and the job ironically gets harder, because signal extraction across a vastly wider pool is the new problem. Strong filters, effective resume review and thoughtful application questions matter more than ever, while keeping the outbound sourcing and referrals running for the quality gems.

Why Solve AI's First Hire Was Ops, Not an Engineer

Steve is technical, ex-Palantir forward deployed software engineer, and was Megan's hiring manager there for five years. The thesis: your first 10 to 15 hires are the architects of your company culture, setting the standard on communication, urgency, ownership and operating through ambiguity. Bringing the ops foundation in first means building the operating rhythm for the hypergrowth phase before it arrives, instead of retrofitting processes two years too late.

The AI Stack Is a Behaviour, Not a Toolset

Solve AI builds on a bunch of underlying LLMs and deliberately avoids tying itself to any one model. Megan's bigger shift is treating AI as a behaviour: if you do a task manually every Tuesday, design the workflow; if you repeat something more than twice, ask what system should exist instead. Her favourite build is a pilot tracker: engineers speak their daily updates into Notion, which flow through against the statement of work and generate automated steerco emails, Slack updates, and a cross-client view of requested product features. She also automated the investor dashboard (inspired by one of Bea's LinkedIn posts). Next up: GTM, where Attio, Clay, Smartlead and an internal app don't yet talk to each other coherently.

Where AI Still Fails: Emotional Nuance

The pilot tracker's auto-generated client email is the cautionary tale. It saved coordination time but lost context, so a human stays in the loop before anything is sent. In enterprise, people buy from people. AI can summarise a meeting, but it can't tell that both sides walked away frustrated, and it won't adjust the tone of the end-of-week update accordingly.

Megan AI vs Cecilia OS

Megan built "Megan AI": an 8am daily rundown that reads her email, WhatsApp and Slack, builds a ranked Notion to-do list and nudges her about neglected relationships. Verdict so far: helpful as a fail-safe, not yet better than pen, paper and held context. Ceci built "Cecilia OS" the same month and landed in the same place: great for catching what slipped and for tracking who owes you a reply, not yet able to keep up with priorities that change hourly.

Hiring in 2026, and the Case for In-Person

Eighteen months ago companies over-indexed on pedigree and logos. Now the vibe is output: leverage, adaptability, speed, and whether you can actually use AI tools, which is increasingly screened in the hiring process. On the perennial in-person vs remote debate, Megan has done both (ElevenLabs fully remote, Palantir in person), and argues that at less than nine months old, trust builds faster in person, and the most important conversations happen in the five minutes between meetings. On headcount, Solve AI is deliberately not in hypergrowth before product-market fit, and will plan hiring ratios linked to AI usage: one person with the right tools instead of three.

Rapid Fire

  • Most overrated hiring metric: time to fill. The easiest to optimise, and the one that quietly lowers the bar by prioritising speed over quality.
  • Most underrated hiring signal: dealing with ambiguity. Plans at Solve AI change weekly, daily, sometimes hourly, and it's the hardest signal to measure.

Frequently Asked Questions

Who is the guest on episode 8 of Top of the Ops?

Megan Standing, Chief of Staff at Solve AI. She previously built and scaled the people and talent function at ElevenLabs from 30 to over 300 people in 18 months, after talent roles at Palantir and Typeform.

Why did Megan Standing leave ElevenLabs?

Not because of a single trigger. She was increasingly drawn to messy, company-wide operational problems rather than purely hiring ones, and Steve, a former Palantir colleague, offered her the chance to be Solve AI's first employee and build everything from zero.

How did ElevenLabs keep its hiring bar during hypergrowth?

Interviewer quality, structured feedback with scorecards and rubrics, accountable hiring managers, founder-reviewed hiring theses with explicit failure modes, founders interviewing every final-stage candidate, and a willingness to slow down rather than make a mediocre hire.

Why was Solve AI's first hire an ops person instead of an engineer?

Because the first 10 to 15 hires are the architects of company culture, and building the operating rhythm for hypergrowth before it arrives beats retrofitting processes later. Steve, the founder, is technical and cares deeply about talent, so the engineering gap was covered.

Is this a guest episode?

Yes. Ceci and Bea are joined by Megan Standing in the show's eighth episode.

Episode Transcript

Chapters
  1. Introduction00:00
  2. From a Decade in Talent to Chief of Staff00:40
  3. What a Decade in Talent Gives You03:29
  4. What She Had to Unlearn06:35
  5. ElevenLabs From 30 to 30008:26
  6. Holding the Bar Without Becoming Cutthroat10:57
  7. Recruiting Before and After a Company Gets Famous13:52
  8. Solve AI and the Ops-First Hire15:46
  9. What Does a Chief of Staff Actually Say They Do?17:48
  10. A Week of Organised Chaos19:58
  11. The AI Stack Is a Behaviour21:27
  12. Where AI Still Fails24:35
  13. Megan AI vs Cecilia OS26:26
  14. Hiring in 202630:15
  15. The Case for In-Person32:36
  16. Rapid Fire35:52
  17. Closing36:49

Introduction

Ceci [00:00]

Welcome to Top of the Ops, the podcast where we have real conversations about what happens behind the scenes of startups. I'm Cecilia, ex-VC at Talis Capital and now FD at PortalOne.

Bea [00:11]

And I'm Bea, ex-VC at Lakestar and now Chief of Staff at Praktika. Today it's not just us. We're very happy to be joined by Megan Standing, currently Chief of Staff at Solve AI, and amongst other things the person who built and scaled the people and talent function at ElevenLabs from 30 to over 300 during one of the most watched hypergrowth stories in AI. Welcome to Top of the Ops.

Megan [00:35]

Thank you so much. I'm really excited to be here and have this chat with you guys today.

From a Decade in Talent to Chief of Staff

Bea [00:40]

So, starting from the headline. You spent nearly a decade in talent: Palantir, Typeform, ElevenLabs. Great logos, actually. Last year you went Chief of Staff at Solve AI. That is a very bold pivot. Take us back to the moment you decided to make it. Was there a single trigger? Did it build over time? What was the thought process behind it?

Megan [01:06]

It's actually a really interesting one, and it definitely built over time. There wasn't this dramatic moment where I decided I wanted to be a Chief of Staff. Across the companies you just mentioned, Palantir, Typeform, but especially ElevenLabs, I found myself increasingly drawn to those really messy operational problems that are more company-based rather than hiring-based. The journey at ElevenLabs in particular, that 30 to over 300, which was in an 18-month period, I was initially hired to come in, take over from the RPO, and own and build the talent function. Very quickly I moved into a broader role within people, which I'd never officially owned anywhere. There were no clear lines at ElevenLabs. It was never "that's not your job", everything is your job. If you see something that's broken, you go, you fix it, you build. I never thought I'd leave ElevenLabs, but Steve, who I'd worked with previously at Palantir, was building Solve, and he said, look Megs, I have this really exciting opportunity to actually be the first employee and build from zero across everything, which I don't think is very typical. Typically you have an engineer as your first hire. That got me thinking about what it is I really like about all the roles I've done to date, and it was the cross-functionality of projects: working with legal on new markets or equity plans, with finance on budgeting, with go-to-market on revenue plans. And then all the visa sponsorship policies, all of these random things I'd never tackled before, where I just needed to go where I was needed to keep the company moving. That's why I took the move. I realised I really liked the ambiguity, the variety, the context switching, and it felt like a natural extension. I don't really believe in job titles very much. I believe more in: go do the thing where you're going to have impact. In my role now that could be compliance, it could be automating in GTM. It's those high-context environments where you can just go do the thing. It gradually built, and I get bored staying stagnant. So what better way to jump in the fire than do something I've never done before.

What a Decade in Talent Gives You

Ceci [03:29]

This is so fun. Job titles, I think, are so reductive in startups, especially really early stage startups. It's more about areas of coverage, which are not defined and always changing and expanding. It's what attracts me personally. But I read your post about the move, and what struck me was how much of people you actually brought with you rather than leaving it behind. So this prompts the question: what does a decade in talent give you in this role that other Chiefs of Staff don't have?

Megan [04:02]

For me, talent gives you that unusual lens on how a company functions. Over the last decade I've spent so many years actually watching how companies scale from the inside out. Palantir, when I joined back in 2017, was in the hundreds, and we went into the thousands during my journey there. You can see, as a company scales, where communication breaks, where you need to put in leadership training, how you shape incentives to scale the behaviour, how you spot bottlenecks. It's judgment in very uncertain situations. In hiring, for example, will someone actually succeed in the role? You never truly know. Interviews are very noisy. References are very selective in the read you're going to get on someone. And people perform differently in different environments: what works at Palantir might not work for ElevenLabs, and might not work for Solve. So what it brought was the ability to naturally think through the people element that affects an organisation's ability to be successful. Because at the end of the day, talent is the machine that determines whether a startup will be successful. You're bringing in the diverse perspectives and you're getting that front row seat on how operational success actually happens. And the side thing I love about being in people and talent is that you get this really unique lens from candidates and employees, where they tell you things they will not tell the executives. You get a really honest understanding of what's working and what's not.

Ceci [05:45]

Something that's not spoken about enough is that you've got to be the ears and eyes of the company, and be the first person to understand when a problem is about to arise. But how do you do that? Very much through relationships, through that communication, and being the person people go and vent to or start raising problems with. I think that's the special sauce. You have to be able to be that person for the company.

Megan [06:12]

I think so. You need to think about all the complexities that come with people. A lot of those problems are maybe a clarity issue, maybe a prioritisation issue, maybe a communication issue. You've got to learn all of those nuances and how to put the pieces together, because all of those things compound and just get worse over time.

What She Had to Unlearn

Bea [06:35]

I'd also give a shout-out to Ashley, the Head of Talent at Praktika, because I've learned this from her. People and talent people have this really good read of people. She would know instantly: this person is not a culture fit, or this person is not going to work out. And 99% of the time she's actually right, which is a read that derives from interviewing thousands and thousands of candidates. It comes from experience. But we were curious about the flip side: did you have to unlearn anything from people and talent when you switched gears into the Chief of Staff role?

Megan [07:14]

One of the things most people and talent professionals do in hypergrowth companies is a little bit of heroics. You say yes to everything, because you want to make sure you're getting all the people in. The thing I had to unlearn was saying no to things that aren't urgent and important right now. My sense of breadth is so much broader now, looking after finance, legal, compliance, security, helping with GTM, and I need to be much more intentional: this is important, but it's not a now piece. I really had to unlearn that people nature inside of me that wanted to say yes to everything and make everyone's lives easier. I still want to do that, but I want to do it intentionally and look at the bigger picture. That was a really steep learning curve, being able to say no because I need to get a SOC 2 done right now, this is the P0, and if we don't get it we're not signing these enterprise contracts. I had to very quickly unlearn being the people pleaser that I am.

ElevenLabs From 30 to 300

Bea [08:26]

Absolutely. I'd like to double click a little bit on the ElevenLabs story. Fun fact for our listeners: Praktika actually uses the ElevenLabs text-to-speech model, it's really embedded in our product, so I've been on the receiving end of what the team has built. But take us back to month one. You're 30 people. What did people and talent even mean at that headcount, and how did that change?

Megan [08:54]

When I joined, I was employee number 28. We didn't have a people function. It was very much: hey Megs, you're just coming in to look after talent. There was no one doing people. We had Victoria doing a really amazing job with all things operations, owning finance, legal, talent, et cetera, but it basically meant there was no separation between functions at that stage. When I came in, it was really about figuring out where we were at: a full analysis of everything related to talent, what's broken, what needs fixing, what's causing us pain, and fixing those very quickly. That could be hiring for senior engineers, then flipping to onboarding, or compensation benchmarking, or visas, relocations, tooling. It was doing a little bit of everything, and very naturally I took over the people elements, because there's such an interesting dynamic between people and talent: once you've found the talent, you want to keep them within the business. The thing that was so interesting is how ambitious the company felt in those first days. Everything moved fast, everything was intense, standards were super high. What I loved at 30, and it continued all the way to 300, is that you'd solve one problem and three more would appear immediately after, because you've grown 3x the headcount you thought you were going to grow. When I joined ElevenLabs, Mati said we were going to be 60 by end of year. We actually ended that year at 130. The pace, all the way from 30 to 300, was about how you preserve that quality. Everyone you meet at ElevenLabs is this very unique type of human: low ego, very high velocity, growth mindset. Trying to keep that as we scaled was the really interesting piece, because culture can change drastically between 30 and 300, but ElevenLabs has very magically kept that culture throughout that period of hypergrowth.

Holding the Bar Without Becoming Cutthroat

Ceci [10:57]

So how do you do it? How do you maintain that bar incredibly high when you're 3x-ing year on year in terms of headcount?

Bea [11:07]

And just to add on to that, for example in Praktika, let's say you hire somebody. Do you give them three months' probation to perform, and then you're very clear about what the expectations are? How do you not end up being a cutthroat culture? That's something that spins out in my brain a lot.

Megan [11:26]

On holding the bar: when you're a really small-scale startup, the recruiter is the quality control mechanism. But when you go through a period of hypergrowth like ElevenLabs, it becomes about how you build out the quality control system so it doesn't break instantly or rely on one human. How we did that was thinking about what's going to scale as the company grows. We spent a lot of time on interviewer quality. How are we going to do structured feedback? How do we have scorecards, rubrics? How do we make our hiring managers more accountable for hiring decisions? We created hiring theses that Mati and Piotr would review, and within those we'd have a failure mode, a success mode, and how we were going to mitigate the failure mode. We were really lucky at ElevenLabs that both Mati and Piotr really cared about talent. That's also what set ElevenLabs up for success. They would interview every single candidate at the final stage. The other thing that needs to happen is you need to be really comfortable with slowing down, which seems very counterintuitive in hypergrowth, but you really need to make sure you're bringing the right people in. If you hire the wrong people, you're going to create much bigger problems than a temporary understaffing. One exceptional hire, in my mind, will outperform four mediocre hires, and one bad hire will definitely compound and affect the team.

Megan [13:00]

One of the things ElevenLabs did to hold the bar without being exceptionally cutthroat is just being really explicit about expectations. Some companies downplay the level of output expected of a hire. ElevenLabs, from day one, before you even join, is exceptionally explicit: this is the bar we operate at, this is what we expect the outcome to be. And then giving really clear, consistent feedback. I know quite a few companies that default to six-month review cycles, so you get feedback twice a year. For us it was about instilling a culture of feedback in the moment, because you have to be very intentional for someone to take feedback on board and change their behaviour. And with any individuals we were thinking of parting ways with as a company, it's giving them the chance to improve. You can't just say, you've done this one thing, you're out. It's: we're going to give you this direct feedback, this is what we need to see, we're going to track it, and let's get there together.

Recruiting Before and After a Company Gets Famous

Ceci [13:52]

Thank you for that. Maybe one last question about the recruiting at ElevenLabs. How does the role of a recruiter change when you go from a small company, where you have to hunt people and convince them to join this crazy new mission, to when the company gets famous and everybody wants to work there? Walk me through that transition.

Megan [14:16]

It's a complete shift within talent. Early on at ElevenLabs it was very much persuasion. You're convincing people to join, convincing them to take a risk on a relatively unknown company. When I joined, it was just at the time of the Series A, so people were only starting to hear about ElevenLabs. But we had a really phenomenal talent team very early on. They were talent mapping, doing multiple rounds of outreach, being creative in how they reached these folks, and doing it 24/7, living and breathing to find the right people for ElevenLabs. Then, when a company becomes known, the main thing that happens is your inbound volume just explodes, and ironically it makes your job so much harder as a recruiter. Earlier you're trying to figure out how to convince person X to join the business. After being more publicly known, it's: who do we actually want to bring on board who understands our mission and wants to work for ElevenLabs for the right reasons, rather than it just being a logo on a CV? That's where you really need strong filters: being really effective with resume reviews, being thoughtful with application questions. The signal extraction becomes so much harder, because your candidate pool has drastically widened, and it's about finding the mechanism where you can still do your outbound sourcing and still get your referrals, but still filter the inbound for those quality gems that do come through.

Solve AI and the Ops-First Hire

Bea [15:46]

I'd like to go to present day, Megan. Solve AI, in your own words: the dinner party version, not the website version. And it's quite interesting that the first person they hired was an ops hire versus an engineer. That says something about the culture and the setup of the company. Do you mind talking us through that?

Megan [16:10]

The background is that Steve is technical. He has over eight years at Palantir, he was a forward deployed software engineer, and he's also been a hiring manager. He cares explicitly about talent and hiring. We actually worked together for five years; he was my hiring manager at Palantir. The reason for me being the first hire is that Steve really resonates with the idea that talent is the most important area of the business, because you have to bring the right people in to set the business up for success. And that's not just hard skills, it's the soft skills as well. Me and Steve align on this: your first 10, 15 hires are effectively the architects of what your company culture will look like. They set the tone, they teach everyone what the culture looks like. We're setting the standard on communication, urgency, ownership, decision-making, and the ability to operate through ambiguity. So the reason for bringing me on as the first hire was to help build that foundation, from a hard skills and soft skills perspective, but also to create an operating rhythm so that when we get to that period of hypergrowth, we're building for the hypergrowth phase, not for the now stage. A lot of businesses bring this person in too late, and by then the company has already built so many processes and operating models that you have to backtrack and rebuild from scratch.

What Does a Chief of Staff Actually Say They Do?

Ceci [17:48]

Yes, so much of every Chief of Staff role is retrofitting and backtracking what's been happening. One of my dreams is joining a company right at the beginning and just sorting out my own systems. But I wanted to ask a question that's one of my favourites. Now you're a Chief of Staff: when somebody asks what you do, what do you actually say, and what do you wish you could say?

Megan [18:10]

I get asked this a lot, because none of my friends are in the tech space. They're like, Megs, what do you actually do? I say that I try to turn ambiguity and complex situations into something that's operationally stable and reliable. And then they say, what does that mean? In practice, I'm trying to reduce organisational drag and make the company operate as effectively, fast and efficiently as possible. And in the age of AI, when everything can happen ten times quicker, it's about using your judgment to figure out what we should build versus what doesn't need to be built, and what still needs a human in the loop. So I say to my friends: keeping the business coherent as we scale. What I'd like to say is that it just means my calendar is an absolute mess, every 30 minutes I'm jumping into something different, and I'm trying to hold all of that context in my head.

Ceci [19:09]

I like to say it's giving ADHD to people that don't have ADHD.

Bea [19:15]

Oh my god, Ceci. I was actually thinking that recently. I've always been, and this is a fact, a very focused person. I never considered myself as having ADHD, not even slightly. I was the person who could focus on something for three hours and not be distracted. And now that has changed a lot, and it makes me a little bit worried about my mental sanity sometimes, because I feel I've developed some type of ADHD.

Ceci [19:43]

I've always been the kind of person who needs about four different windows full of tabs, constantly jumping from one to the other. So the Chief of Staff role is perfect for me: I've got People, Product, just jumping around. It's really not helping, actually.

A Week of Organised Chaos

Bea [19:58]

I'd be curious to ask what your operational cadence is. How do you make sure you block time in the diary and say, okay, I'm going to have these four hours on Thursday where I just do focus work? Tell us a little about how you organise your week.

Megan [20:13]

My week is organised chaos, but I do like structured blocks. This week, for example, I have a block because I know I need to do our 409A. I know I need to review some AI tooling for recruiting. I've got metrics reporting for our investors, and some finance work. Everything is blocked, but I give myself enough flexibility that if something super important comes in, I can shift. Today we're onboarding a new client and they said, Megs, I need you to do this security review now. So, cool, I park this and jump into that. I like having structure because there are so many competing priorities right now. We're less than nine months old as a business, and with what we've raised so far, the clients we're onboarding and the hiring we're doing, I need some structure, but with enough flexibility to shift things around. And we operate exceptionally transparently as a business. I'm an outlier today because I'm at home, but typically we're in office, which makes everything so much easier and lets us move quickly.

Ceci [21:20]

Organised chaos. It's the perfect way to put it.

Megan [21:23]

Yeah. Everything is colour coordinated and blocked, but chaotic.

The AI Stack Is a Behaviour

Ceci [21:27]

It seems to be impossible to do a podcast in 2026 without talking about AI, so here we are. Can you talk us through your AI stack right now? What do you use, and what do you use it for?

Megan [21:38]

I actually do a lot of the building for our own platform, which sits on a bunch of different underlying LLMs. It really depends on the use case, and models are evolving so quickly that we don't want to tie ourselves to anyone right now. But I think it's not actually about the stack. I've had a bigger shift: I don't really care about the tool, it's more about the behaviour. And not just my behaviour, but everyone's on the team. I'm trying to get everyone to think: am I doing this task manually every single Tuesday? If so, let's use AI to design the workflow and optimise it. If you're repeating something more than twice, I'm instinctively asking myself and everyone else: what's the system we should be building instead? I think of it as a behaviour rather than the underlying tech stack.

Bea [22:28]

Would you give us a practical example of something you've automated that saved you a huge amount of hours?

Megan [22:33]

Something I've recently automated: we're now working on a bunch of pilots, and we didn't have a single source of truth, so it was really hard to understand the current state of those pilots. What I've built is a pilot tracker. Within Notion, on a daily basis, our engineers speak to Notion and give all their updates for the day. That pulls through to the pilot tracker against our statement of work. It then creates automated steerco emails at the end of the week, automated Slack updates, and tracks progress against the plan: wins, blockers, pain points. That's saved us so much time, because we have a single source of truth, and it also pulls all the product features from those notes, so we can see that across X clients, they're requesting Y products and features. That's quite a big one for me. The other one, Bea, actually comes from a post you did on LinkedIn. I've now automated our investor dashboard, which is a massive time save: I'm not having to do Excel spreadsheets, and I have very clear, concise dashboards for hiring, customers and finance. Those are the two I'm particularly proud of recently. And the next big project is GTM. I'm going to be looking at how we do better automations, cleaner processes, and getting all the systems talking to each other. At the moment we have Attio, Clay, Smartlead and our own internally built GTM app, and they're not all talking to each other as consistently or coherently as I'd like. So that's my next meaty project: how we use AI in the most effective way to improve go-to-market, our lead lifecycle, and reporting.

Where AI Still Fails

Ceci [24:35]

Still on the AI topic: where does it just not work? Where do you still need the human touch?

Megan [24:42]

It's really interesting. The pilot tracker I just mentioned, the one place it broke is where you need emotional nuance. The tracker is fabulous, and I was experimenting with the steerco email it generates at the end of the week to give the client all the critical key updates. It failed, because it lost some of the context we were trying to get across when it pulled the information through. So what I've done is it still generates the framework of the steerco email, which operationally saves a huge amount of time and manual coordination, but we keep a human in the loop before we click send on that email each week. The reason is that, working in enterprise, most of it is about relationships. People buy from people, and you need to make sure everything you send is accurate. The AI can summarise to some extent, but it can't fully understand stakeholder sentiment. If it's recorded a meeting where both sides walked away feeling a bit frustrated, the AI is not going to capture that. But having been in the meeting, you can change the tone of the end-of-week update you're sending to the execs. That's where I'm not using AI: where emotional nuance, tone or framing needs to be maintained to keep trust. A lot of that at the moment is interactions with our clients.

Megan AI vs Cecilia OS

Bea [26:26]

I'd be interested to ask whether you use any "office of the Chief of Staff" type of AI automation, I don't even know how to define it. I know some Chiefs of Staff have an automated planner: Monday morning it hits their inbox, this is the plan for the week, with self-adjusting planning of their own work. It's something I don't have, because I think I always have my priorities pretty clear, but it would be interesting to try. Do you use anything like that, or is everything still in your head?

Megan [27:05]

This is really interesting, because two weeks ago I built Megan AI. Do I find Megan AI as useful as just holding context in my head? Not yet. No. What my Megan AI does is run at 8am every single morning. It gives me a rundown of what my day looks like and breaks down all my top priorities. It gives me an open task list: it reads my emails, my WhatsApp and my Slack, and it goes into an automated Notion to-do list for the day, ranking things by priority. It's definitely helpful in some respects. It also gives me relationship nudges: you got this email from X person, you haven't checked in with them in three months, you should reconnect today. That's helpful to the extent that it might highlight something I've forgotten. But, and I don't mean this in an ego way, I have quite a good mechanism of pen and paper, old school style: this is super important for this week or this month. I love crossing things off and holding that state in my head. So Megan AI is helping in some respects, but I'm not fully bought in yet.

Ceci [28:21]

So interesting that you said that, because last week I built Cecilia OS. I thought, okay, let's give it everything. There are some things that are quite useful: it crawls Slack, emails, notes, my Notion to-dos, which it can update and move around. That's super useful. But my priorities, and the things I move my attention to, are constantly changing, and it's not able to keep track of that as fast as I'd like. I kind of wanted it to read my brain, and we're not there yet. What I really like about it is two things. One, it's a fail-safe to catch everything I might have missed in the day, or didn't save on Slack. And the other is a way to remind myself who owes me what, because that's something I've always struggled to keep track of. Now I have an agent running around saying, that person hasn't replied to you, that person hasn't replied to you, and pulling it into a tracker. I don't want it to act on anything, because I know how long it should take somebody to get back to me on a specific thing, but I just need to see it. So that's where I'm at. It's been one week. We're still getting to know each other.

Megan [29:35]

We'll see how it goes. The thing I also struggle with is that, as I said, we're mostly an in-person organisation, so sometimes that context lives in conversations rather than async on Slack or email. That's one of the nuances of why I'm not there just yet and need to make some tweaks to the workflows.

Bea [29:55]

I also think it's not a bad thing that our brains still function better. It makes me hopeful for humanity.

Ceci [30:04]

And I actually really like, I'm really proud of my memory. I have a really good memory.

Bea [30:10]

I don't. I have the memory of a goldfish. I have to write down everything, which is why AI is really good for me.

Hiring in 2026

Ceci [30:15]

We wanted to touch upon what hiring is like in 2026. Every hot company is hiring at the same time, and they're all looking for the same archetypes. Can you tell us what it's like to be in the talent market right now versus 18 months ago?

Megan [30:32]

18 months ago, I feel the majority of companies were probably still over-indexing on pedigree or logos. Now the vibe is definitely much more output-orientated. There's so much more focus on what someone's leverage and impact can be. Are they adaptable? Can they move fast? Can they operate with ambiguity? Can they use AI tools effectively? There are so many more people screening for AI capability throughout the hiring process. That's quite a dramatic shift from where we were 18 months ago.

Ceci [31:06]

Talking about your journey at Solve AI: you were employee number one and you obviously hired a team. You do all your work, you make sure everybody is who you want them to be. How do you make sure the hire succeeds?

Megan [31:20]

A lot of a hire succeeding relates to making sure every candidate has a positive employee lifecycle and feels valued at every single point. A lot of that comes down to being really open, honest and transparent with the level of feedback, making sure people get the positive feedback as well as the constructive. I've definitely been in environments where you're moving so quickly that you forget to give the positive feedback, because you're right onto the next thing. It's about figuring out, for each person, what motivates them and how we help them succeed. A lot of it marries to the types of individuals you're hiring: you're thinking about the personas that will be successful in your organisation, so naturally they should be quite motivated, resilient, et cetera. But you have to give them the foundations to set them up for success, like a super strong onboarding. And I'm a really big believer that people stay in companies because of the people they work with. If you work with great people, and you spend most of your time at work, you're going to stay if you're aligned to the mission and everything else. So it's about being thoughtful, bringing people together, and keeping people engaged and happy with the work they're doing.

The Case for In-Person

Ceci [32:36]

Let me ask you one last question, and it's something that is super divisive in the startup world: in-person versus remote. Why are you an in-person believer? As an anecdote, both Bea and myself work remotely, so I'd love you to convince us of the strength of in-person.

Megan [32:58]

Just for context, I've worked remote previously. ElevenLabs was fully remote, and Palantir was in person and then went remote post-COVID. Why it's so important for us now, and why we're focusing on an in-person environment at Solve, is that we're at this super early stage. We're less than nine months old, and we're planning to go through a period of hypergrowth. A lot of that building happens through the speed of in-person communication, and through the trust we formulate. I'm still a massive believer that you can build high-trust relationships much faster in person than over a VTC. We're operating in a period of ambiguity where quality and alignment really matter, and some of our most important conversations happen in five minutes in a meeting, or in between meetings, not in scheduled calls. I know plenty of people who do incredible work remotely, and I think it's less about in-office versus remote, and more about what's going to set us up at this stage, when we're so new and formulating as a team. It gives us high trust, velocity, clarity. And we have new people joining: our first front-end engineer joins on Monday, and being in person, being able to get them up to speed and give them context, is invaluable. I don't know whether that means we'll stay fully in person forever, but at this stage it's really helping us formulate.

Bea [34:25]

I have a last question before the rapid fire. Do you think you're going to experience hypergrowth on the people and headcount side, as much as you experienced it at ElevenLabs? Or, now that we're in an AI world, will you be more intentional with hiring and use AI heavily, as opposed to scaling to a 200-person engineering team? What's your opinion?

Megan [34:50]

We're definitely trying to be very thoughtful on headcount. We're at this very unique phase where we haven't completely gained product-market fit, so at this point in time we're not going to go through hypergrowth. We'll be very intentional, utilising the resources we have to sign the pilots and sign the enterprise deals. We're very different from ElevenLabs, for example, because we don't have B2C. We're purely B2B, and enterprise SaaS is hard. When we crack the market, we'll think about hiring and ratios linked to AI usage. We did a lot of this previously at ElevenLabs and Palantir: if we have X amount of contracts, we'll need X amount of legal to review them, X amount of AEs to hit these ARR targets. So to answer your question more succinctly: it's definitely something we'll be cognisant of. Can we bring in one person and utilise these tools to be effective, rather than needing three people to do the same job?

Rapid Fire

Bea [35:52]

That makes a lot of sense. Let's go into the rapid fire. The first one, and sorry to always bring it back to hiring, but we're so curious: most overrated hiring metric?

Megan [36:06]

I would say time to fill. It's one of the easiest metrics to optimise, but it's one of the metrics that lowers the bar considerably in terms of talent quality, because you prioritise time over quality.

Ceci [36:19]

Most underrated hiring signal?

Megan [36:22]

The most underrated for us right now is people who can deal with ambiguity. A lot of companies have quarterly plans, yearly plans. We literally do monthly planning and then it changes the next day. Most of our plans change on a weekly basis, a daily basis, an hourly basis even. So the most underrated signal is dealing with ambiguity, and it's hard to measure.

Closing

Ceci [36:49]

Well, this is the end of the episode. Thank you so much, Megan, for being here. We've learned a lot about what you're doing now, especially about hiring and what a hiring background can bring to a Chief of Staff role.

Bea [37:03]

We loved having you on. And for everybody listening, please go and follow us on YouTube and Spotify, hit subscribe, give us some love, and thank you so much for listening. See you next time.

Keep Listening

8 Ways the Chief of Staff Role Fails (And How to Fix It)
Two Chiefs of Staff name the eight ways the role quietly goes wrong, with the honest fix for each.
Clara Ma, from Hugging Face CoS to Ask a Chief of Staff Founder. Are Chiefs of Agents the New Chiefs of Staff?
Hugging Face's first CoS on the rise of the "Chief of Agents", why the role can't be automated away, and her 30-60-90 day playbook.
Alejandra Otero, First Principles and Operational Excellence at ElevenLabs
More from inside ElevenLabs: first-principles thinking and operational excellence during hypergrowth.