The Twenty Minute VCElevenLabs: Building an AI Sales Machine & Why We Set a 20x Sales Quota
CHAPTERS
- 0:00 – 2:47
AI-first CRO mindset: distribution, not just closing deals
Carles and Harry frame why many traditional CRO playbooks are falling behind. The core shift is moving from deal-by-deal execution to building durable distribution systems, with AI embedded throughout the revenue engine.
- •Market dynamics have changed: faster deal cycles, more competition, more experimentation
- •Modern CRO remit: design distribution, not just manage a pipeline
- •AI should increase leverage across the whole GTM system
- •CRO role is about building “revenues of tomorrow,” not just today
- 2:47 – 4:15
Why most AI outbound/AI SDR tools fail (and what “humanly” automation looks like)
Carles rejects the common approach of AI SDRs blasting templated messages. He argues most tools treat outreach as transactional, causing response rates to collapse, and outlines what a genuinely personalized, channel-aware agent should do.
- •Most AI outbound tools assume everyone wants email/LinkedIn outreach—buyers hate it
- •Outbound response rates have cratered; AI-generated messaging is easy to detect
- •The right model: infer buyer preferences (email/events/calls) and tailor accordingly
- •Outbound isn’t dead—“outbound is dead unless you do it with humans or humanly”
- 4:15 – 5:47
Inside ElevenLabs’ revenue agents: inbound SDR, RFP scanner, and CS copilot that closes deals
Carles describes the internal AI agents ElevenLabs built and why they work: they augment humans with drafts, context, and proactive recommendations. The system learns from what gets sent and customer responses, improving personalization over time.
- •Internal push to hire engineers to build revenue agents, not buy generic tools
- •AI inbound SDR delivering strong results
- •AI proposals manager scans the web for RFPs/RFIs and suggests scoring/responses
- •AI CS manager drafts proactive, contract-aware, multilingual customer emails
- •Closed-won impact: AI-assisted CS outreach has already generated revenue
- 5:47 – 6:45
Will AI shrink sales teams? The 50% productivity target and “smaller, elite, highly paid” orgs
Carles expects AI to materially improve per-rep productivity and reshape team sizing. His goal is 50% productivity improvement so ElevenLabs can hire fewer people while paying top performers more, including commission credit for AI-assisted closes.
- •Target: 50% productivity improvement across revenue org
- •Preference for smaller teams with higher compensation and talent density
- •AI-closed upsells/contracts still get paid like a human closed them
- •Industry implication: slower headcount growth, more concentration in top talent
- 6:45 – 9:05
The 20x quota + comp philosophy: challenge, accelerators, and why huge commission checks are rational
The discussion moves into quotas and commissions, including the headline 20x quota model and cases of reps hitting full-year quota by February. Carles explains why aggressive quotas can be fair when paired with accelerators and a value-based view of revenue’s impact on valuation.
- •Anecdote: two reps hit full-year quota by February
- •Quotas should be challenging but adjustable if truly unfair
- •Accelerators matter: over-performance should unlock meaningful upside
- •Valuation lens: each $1M revenue can imply massive valuation uplift—paying big commissions is rational
- 9:05 – 11:28
Comp mechanics that avoid bad behavior: no pilot commission, spiffs with caution, retention/expansion pay
Carles outlines the practical structure: baseline commission, tiered accelerators, and product-specific spiffs—while warning that too many incentives distort behavior. He details why pilots don’t earn commission, and how retention/expansion and strategic accounts are compensated.
- •Commission starts at ~5% with accelerators (1.1x, 1.2x, 1.3x, 1.5x, etc.)
- •Spiffs can help focus but can also create mis-selling if overused
- •No commissions on pilots (don’t add valuation; can drive wrong incentives)
- •Commissions on expansion/retention; strategic accounts can earn for up to two years
- 11:28 – 14:18
Hunters vs farmers: making Customer Success a revenue function (and why “services-only CS” fails in AI)
Carles argues that in an AI world where competitors can emerge quickly, Customer Success must drive expansion and retention, not just satisfaction. He contrasts older-era views (e.g., CS as “bullshit” or pure paid services) with today’s need for community, trust, and speed.
- •Hunters can create pricing/value chaos if unmanaged on complex accounts
- •CS must be a money-generation function: expansion, cross-sell, retention
- •AI competition compresses moats—deep post-sale engagement matters earlier
- •Charging for services can be appropriate, but “transactional everything” weakens long-term retention
- 14:18 – 19:53
“One market at a time” is outdated: parallel bets and hiring veteran sellers to compress cycles
Carles challenges the classic VC advice of sequential geographic expansion. He advocates running multiple GTM bets in parallel and hiring experienced sellers (20+ years) who can go straight to C-level relationships and shorten enterprise cycles.
- •Sequential market expansion is too slow when competitors proliferate rapidly
- •Parallelism: treat GTM like venture—many experiments to find a few big winners
- •Experienced sellers can bring relationships and accelerate enterprise access
- •Culture fit isn’t about age; it’s about hunger, alignment, and shared drive
- 19:53 – 33:59
What didn’t work: entertainment studios, shifting ICPs, and the “India reset” after premature verticalization
Carles shares failed or suboptimal bets and how ElevenLabs corrected them. He explains why selling directly to major entertainment studios stalled, why enabling creator platforms worked better, and how an India segmentation attempt hurt performance until they rebuilt the approach from scratch.
- •Early bet on big entertainment studios didn’t convert fast enough
- •Pivot to media creation platforms improved traction
- •Agentic systems became a major growth driver after early investment
- •India mistake: verticalized too early with too-small team; revenues dipped; reverted to horizontal and rebuilt pipeline
- 33:59 – 38:09
Pipeline construction as portfolio construction: liquidity + whales, government as a long-horizon bet
Carles describes designing pipelines the way a VC constructs a portfolio—balancing quick wins with big strategic opportunities. He uses government sales as an example: long cycles and hard entry, but sticky outcomes, societal value, and long-term defensibility.
- •Weekly “pipeline construction” ritual to design the right mix per market
- •Liquidity deals keep rep confidence; whales drive outsized outcomes
- •Government vertical: hard to sell into but sticky once landed; also mission-driven value
- •Experimentation mindset makes traditional forecasting precision unrealistic
- 38:09 – 40:24
Brand as a sales-cycle accelerator: becoming the ‘safe choice’ in enterprise procurement
The conversation turns to how brand materially reduces enterprise friction. Carles argues brand is decisive—procurement wants safety—and cites OpenAI, Anthropic, and Cursor as “blue chip” examples that shorten cycles and reduce buyer fear.
- •Brand dramatically shortens enterprise sales cycles; “no one gets fired” buying safe brands
- •Procurement and IT teams optimize for perceived risk reduction
- •Cursor cited as a standout enterprise brand execution
- •Goal: replicate blue-chip status for ElevenLabs globally
- 40:24 – 48:16
Scaling the sales org: radical transparency, high standards, and spotting obsessed talent fast
Carles explains how ElevenLabs hires and manages for intensity: explicit expectations, internal transparency, and performance visibility. He shares how leaderboards work, what he looks for in interviews, and why he can decide in ~20 minutes whether someone fits.
- •Hiring message: high expectations, long hours, full commitment—filters for fit
- •Internal transparency: per-rep quota attainment shared openly via automated reporting
- •Performance evaluation goes beyond revenue: strategic account building and pipeline actions matter
- •Talent signal: energy, sharp thinking, depth, preparation; rapid “pitch me” prompts reveal it
- 48:16 – 49:10
Experimentation over budget limits: small tests earn scale; when to build vs buy internal tools
Carles argues budget is rarely the constraint—proof is. He prefers lightweight experiments over large upfront commitments, and discusses when it makes sense to build internal software (e.g., CRM) versus relying on standardized tools.
- •Rule: fund small experiments quickly; don’t commit massive resources without evidence
- •Constraint is lack of validated learning, not dollars
- •Selective “build your own software” future—yes for core workflows like CRM, no for commoditized suites
- •Sales tooling stack: Gong + Salesforce; interest in more customized internal systems
- 49:10 – 55:29
Brand-building and events: F1 sponsorship logic, executive dinners vs conferences, and designing non-salesy summits
Carles details how ElevenLabs and Revolut approached major brand bets like F1 sponsorship and owned events. He shares why executive dinners outperform trade shows, how to measure influenced vs closed revenue, and what makes customer events compelling without feeling like a pitch.
- •F1 sponsorship: negotiated multi-team options; chose fit + narrative alignment
- •Preference: premium branding on weaker team can beat secondary branding on top team (exposure + creative freedom)
- •High-ROI enterprise tactic: curated exec dinners that create FOMO
- •Conferences often poor ROI; owned events can be measured via influenced/closed revenue and must avoid being overly salesy
- 55:29 – 1:06:55
Partner ecosystems done right: CVC strategy partners, incentives, and why partnerships take time
Carles explains ElevenLabs’ approach to partnerships, especially leveraging corporate venture capital as strategic distribution. He emphasizes incentive alignment, time horizons, dedicated resourcing, and metrics evolution from pipeline signals to revenue outcomes.
- •Strategy partners: CVC-backed relationships create internal champions and distribution paths
- •Incentive alignment: investment tied to revenue commitments; consequences can include buyouts
- •Ecosystem maturity requires dedicated teams and patience—no quick 20% revenue lift in two quarters
- •Partnership measurement: start with SQL/SQO volume, then graduate to revenue metrics
- 1:06:55 – 1:25:37
Operators as investors + quick-fire: balancing focus, what Carles optimizes for, and the next foundation-model wave
Carles argues operators should invest if they can remain the hardest worker and be genuinely helpful beyond money—especially on GTM and hiring. In the closing quick-fire, he shares personal goals, stressors, leadership admiration for ElevenLabs’ CEO, and his excitement about new foundation model entrants and M&A.
- •Operator-investor model works when it reinforces learning and you maintain excellence at the day job
- •Best way to invest: be helpful—GTM feedback, hiring, customer calls, intros; check size matters less than support
- •Quick-fire: 2026 happiness = ElevenLabs at $1B revenue + multiple unicorns in fund; biggest challenge is time; escape is gardening
- •Excitement: next wave of foundation model companies and consolidation via acquisitions; preference for Anthropic over OpenAI at stated valuations
