In GTMBA’s Revenue Leader Series, we give a voice to early & growth stage GTM leaders building their company's Go-to-Market motion. In today’s spotlight, we sat down with Carles Reina, VP Revenue, International @ ElevenLabs, and solo-GP at Baobab Ventures. Carles was a seed investor in ElevenLabs, joined as the first GTM hire and helped scale from $0M to >$330M ARR. An early Uber operator in Europe turned startup builder and investor, Carles has strong views on how modern AI companies should scale: technical founders hire sales operators early, revenue teams should be rebuilt around specialized AI agents, and distribution - not just product - is what determines who survives. He’s equally outspoken about Europe’s startup rise, the future of partner-led growth, and why money alone is no longer enough in venture.
Julian: Carles, you’re GTMBA’s first non-US guest and are a loud champion of the European start-up ecosystem. Are Europeans too focused on work-life balance to build a successful ecosystem?
Carles: I’m passionate about Europe because we’re finally seeing the startup ecosystem become self-sustaining. Ten years ago, building a start-up in Europe wasn’t “sexy.” Now, you have a growing group of ambitious people who don’t want a traditional corporate path. The dynamic is continuing to bifurcate. One group optimizes for the stability of a 9-to-5 life, and another is drawn to the intensity and ambition of startup building.
The shift is more than a result of a few breakout companies. Companies like Revolut became landmark success stories, but the bigger unlock was the normalization of secondaries, acquisitions, and employee liquidity. Once founders and employees can point to real financial upside, more people start to believe they can do it too. That creates a snowball effect: more talent enters startups, more wealth is created, more angel investing happens, and the ecosystem compounds.
My biggest pet peeve is when VCs take too much credit as benefactors of founders, rather than participants in helping build a compounding start-up ecosystem. I’ll hear VCs claim: “I bought houses for my founders.” Rather, the best founders make a smart decision to derisk by selling some shares, and investors get the opportunity to increase ownership.
Julian: How is selling outside the US different from in the US? Walk us through your market entry approach in non-US markets.
Carles: Our motto is “build globally, deploy locally.” Global markets do not behave like the US, and founders underestimate how much local nuance matters.
For example, in markets like Latin America and Japan, prospects may not tell you “no” directly. It’s easy to develop happy ears if you’re not in tune with local cultural norms and waste months misreading politeness as momentum.
Another example is communication channels. In the U.S., most B2B conversations happen through email and Slack, but in many regions, business conversations frequently happen on WhatsApp. Deals progress through ongoing chat threads rather than long email chains.
Successful international expansion requires local talent and constantly traveling for repeat in-market exposure.
Julian: GTMBA believes more bankers/consultants/MBAs should go into AI/start-ups/GTM. You hated working in banking because it rewarded compliance over creativity. What’s one anti-process principle you carried from that experience into your unicorn runs?
Carles: I reject risk-averse, hierarchy-driven thinking. In banking, everything had to be done in a prescribed way: high quality, highly structured, heavily process-driven, and deeply hierarchical. This mindset is a liability inside startups. Too much process kills both creativity and innovation.
I also want to distinguish between startup reality and “startup theater”. At Uber, I saw how easy it is for teams with large amounts of money to confuse spend with progress. In the AI market today, when companies have enormous access to capital and compute, they can brute-force problems instead of finding more creative, efficient solutions.
In the early days of ElevenLabs, by contrast, we had to compete with the large AI labs and didn’t have massive capital or compute resources. This forced us to be cost-conscious - optimize our model training approach and build the business with tighter constraints. More money often makes teams lazier because it encourages the easiest path rather than the most inventive one. On GTM, we ran a large number of experiments to determine our ICP, accepted that most would fail, and optimized for the few that generated outsized outcomes.
Julian: You notoriously called Founder-Led Sales dumb. Elaborate.
Carles: The context matters. I’m specifically referring to technical founders and I think in two stages. Before you have a real product - your job as founder is to recruit design partners. Those early users or customers are helping shape the product, and that relationship is fundamentally different from a repeatable sales motion. But founders often confuse early design partners with a repeatable sales engine.
Once you have several design partners giving feedback, founders should bring in a professional seller much earlier than most VCs recommend. Sales is not the founder’s core craft. The first GTM hire should be someone with similar energy and ambition, but who is wired to run experiments across ICPs, messaging, qualification, and feedback loops. That person’s job is to turn isolated early traction into actual commercial motion.
No investor would tell a GTM-specialized cofounder to suddenly become a world-class engineer, so why do investors keep telling technical founders they should also become expert salespeople? VC advice is often less about company-building and more about investors trying to derisk their own position.
Bring in someone who is both entrepreneurial and heavily incentivized. I like equity tied to revenue milestones and clear time-based checkpoints, because it creates the kind of challenge-oriented upside that early sales talent is attracted to.
When you are closing deals consistently every week, you are starting to move into actual sales traction. For AI-native businesses, you don’t have true product market fit until $100M ARR. Smaller businesses can be good businesses, but in this market, subscale companies remain highly vulnerable to larger model providers and distribution-heavy incumbents.
Julian: You said the ideal early seller “has zero ego, runs experiments, and thinks like an engineer in search of ground truth.” How do you test for that in interviews?
Carles: I look for flexibility, sharpness, and the ability to think under pressure. I put candidates into progressively more uncomfortable, dynamic situations. For example, I will first ask the candidate to pitch ElevenLabs, or walk through how he/she would manage 80 or 100 accounts. Then keep changing the conditions midstream. The goal is to see how they adapt, how creative they get, and whether they can iterate in real time and pause/reflect before launching into responses.
Salespeople are often trained to talk their way through anything, so part of the interview process has to be designed to expose whether someone is BS’ing or actually thinking. Past execution still matters - top-performer history, prior outcomes, evidence of having actually won - but the deeper test is whether the person can operate like an experimentalist instead of a script-reader.
Julian: In 2024, before agents were cool, you wrote a “Revenue Vision” to 10x internal GTM rep productivity at Eleven Labs using AI agents. What did you launch?
Carles: In January 2024, I pitched re-architecting our GTM team and allocating engineering headcount directly to GTM to build agents. We mapped the entire revenue workflow and built specialized agents for those specific moments. Today we run several: AI SDRs, an AI coach, an AI CSM, and more.
Our typical CSM manages 100-200 accounts. In reality, that means most of the attention goes to the largest customers and the long tail gets neglected. That’s dangerous, because many of our biggest customers didn’t start big. They started as self-serve users spending $20K-$30K, and eventually grew into 7-figure accounts. If you don’t engage those accounts early, you miss the outliers.
The AI CSM solves that problem. It analyzes account activity, drafts engagement emails, and keeps those accounts warm so the team can identify expansion opportunities earlier.
We trained the agent based on actions of one of our strongest CSMs - feeding Gong calls, email history, and account data - so it learns how top performers actually communicate with customers. The accounts are still fully owed by the CSM, but the system acts as an extension of the account owner.
Our agents add the most value in supporting the long tail and removing operational work - analyzing data, drafting outreach, preparing things like QBRs. Our AI CSM is optimized to send 20-30 high quality emails per day. .
Over the next year, it’s likely we’ll start seeing agents autonomously closing simpler deals in high-velocity segments.
Julian: Why is building a Partner Sales motion so important today?
Carles: Partner sales is central to revenue quality and long-term defensibility. Many companies over-index on large direct deals. Partners can be incredibly valuable to build:
Resilient Revenue: If too much revenue is concentrated in a small set of large accounts, the business becomes fragile: one or two churn events can punch a hole in the company. Sustainable growth requires a mix of customer types (enterprise, strategic, mid-market, corporate, PLG), and revenue sources (direct - outbound/inbound, resellers, GSIs, and more).
Brand Ubiquity: If you want to build a generational company, people need to encounter your brand constantly - one way to do this is to embed your product into industries, workflows, and geographies you cannot efficiently serve alone.
Localization: Partners also become especially useful when entering markets where you do not yet have local entities, regulatory fluency, or language capabilities.
One cautionary tale: some partners mainly want logo association rather than real execution. So partner motions need to be tested hard and measured by actual output.
One of the most underutilized partners is government agencies. While many startups focus on resellers and GSIs, government agencies are strategically important for generative AI platforms. The payoff is slow, likely 18-24 months - but if executed well, it can create deep entrenchment in a market.
Julian: You run a solo-GP VC fund on the side called Baobab Ventures. Why is this GTM Operator to VC trend becoming so popular?
Carles: Money has become a commodity. If founders can get capital from multiple places, then the deciding factor becomes whether that capital comes with real leverage. The new generation of operator-investors wins because they can help companies do something specific and important: turbocharge go-to-market, improve hiring, sharpen execution, or navigate difficult markets. Founders increasingly want that operational edge, not just a name on the cap table.
This is part of a broader evolution in venture. Traditional firms built around access to capital and pattern-matching no longer have the same natural moat. a16z recognized this early and built a wider ecosystem of support around capital - operating partners, growth teams, and deeper involvement in adjacent arenas like government and policy. Solo GPs are an extension of the same trend, but with even sharper specialization.
The edge for the true operator-investor is recency. Large-firm operating partners may have relationships and prestige, but they are often no longer “in the game” in the same way. They may not understand what it feels like right now to open Japan, negotiate a live deal in a fast-moving AI market, or build GTM amid today’s constraints. The solo GP advantage is currentness: speed, specificity, and firsthand relevance.
For additional resources on Carles’ journey and approach to building GTM teams, check out 20Sales or Scaling Europe.