Building an AI Sales Machine, Why We Set a 20x Sales Quota | Carles Reina

Why does it matter? Because the AI sales playbook most companies are betting on is actively destroying their pipeline.

Carles Reina built 11 Labs from zero to $350M ARR — and his core message is a direct rebuke of how most companies are spending their go-to-market budgets right now. AI SDRs are burning prospect trust, sequential market entry is a death sentence in a world where 100 competitors appear in a month, and Customer Success teams structured around satisfaction scores are handing expansion revenue to hungrier rivals.

• Outbound email response rates have collapsed below 0.01% — AI spray-and-pray is training your best prospects to ignore you permanently
• The only AI sales stack that actually closes deals is human-in-the-loop: AI drafts, human approves, system fine-tunes on what converts
• A 20x quota with steep accelerators is a culture mechanism, not a punishment — and commission should pay out on AI-closed deals at full human rates
• Corporate VCs with contractual revenue commitments attached to their allocation are a stealth distribution weapon that no outbound motion can replicate

AI SDR outbound is dead — and the companies still buying these tools are paying to poison their own pipeline

The response rates on outbound emails have dropped to less than 0.01% — the lowest at any point in history. Reina has tested a large number of AI go-to-market tools and reached a blunt conclusion: they don't work, because "they see everything as a transaction."

The problem isn't AI itself. It's that these tools treat every prospect identically — blasting messages without understanding whether someone prefers email, events, or phone calls. Prospects can instantly perceive a transactional mass-send, and LinkedIn inbound messages have skyrocketed to the point of becoming noise. The window isn't just closing; for many companies it's already shut.

"Outbound is dead. Outbound is dead unless you do it with humans or unless you do it humanly."

The compounding damage matters here. Every spray-and-pray sequence that lands in a prospect's inbox doesn't just fail to convert — it trains that person to filter out your brand. Billions are being allocated to AI SDR tools that are actively degrading pipeline quality quarter over quarter. The companies that stop now will have a measurable advantage in 18 months over those that keep running the playbook.

The AI sales stack that actually closes deals is built on a human in the loop, not a human out of the picture

11 Labs is already generating revenue from AI-assisted sales — and the reason it works is precisely because it doesn't feel like AI. Reina's team built an AI Customer Success manager that lives in the background of email: it scans all customer data, pricing tiers, contract details, and proactively drafts responses. The human CSM arrives in the morning to a queue of AI-generated drafts, edits them, and sends. Every sent message and every response gets stored to fine-tune future drafts — so each customer receives a progressively more personalized tone, and multilingual customers get outreach in their own language.

"That works and that has closed deals for us already. We generate money from that AI customer success manager fundamentally because it is humanly."

On top of that, an AI proposals manager scans the web for RFPs and RFIs, scores them, and surfaces the best opportunities automatically. The productivity target Reina has set internally is 50% improvement — not to double headcount, but to hire fewer, better-compensated people. Any upsell or contract closed by an AI agent pays commission at the same rate as a human close. The incentive structure is deliberately designed so reps want the AI to succeed.

A 20x quota isn't sadistic — it's the mechanism that creates a team worth going to war with

Two 11 Labs reps hit their full annual quota by February. Reina posted it in Slack and called it "Mount Olympus of sales." To the obvious challenge — didn't you just prove you set quota wrong? — his answer is that high quota combined with steep accelerators is a deliberate design choice, not a miscalibration.

The math that drives his conviction: every $1 million in revenue signed, by anyone — SDR, AE, CSM, or engineer who lands a contract — generates $33 million in additional company valuation. From that frame, a million-dollar commission check is the happiest outcome imaginable.

The accelerator structure starts at 5% base commission on anything sold. After quota, multipliers kick in at 1.1x, 1.2x, 1.3x, 1.5x and beyond — plus product-specific spiffs when the company wants to push a particular offering in a given quarter. One hard rule cuts across all of it: no commissions on pilots. Pilots don't add to company valuation, so engineers and researchers see no equity boost from them — paying sales commission on a pilot would misalign the entire organization. Commission triggers only when an annual or multi-year contract is signed, and expands on renewal and upsell for strategic accounts over a two-year window.

Sequential market entry is the worst advice in AI — treat go-to-market like a venture portfolio

"This is the fucking advice that VCs have been giving for many years" — go deep in one market, prove it, then open the next. Reina calls this counterintuitive at best, catastrophic at worst in an AI environment where 100 competitors can emerge within a month of you getting traction.

The replacement framework is explicit: go-to-market is portfolio construction. Test a hundred things to find the three, four, five that actually perform. Balance high-liquidity markets that generate quick wins against long-cycle strategic bets. Enter multiple simultaneously with a reasoned thesis for each — so you can track whether market conditions are changing and evolve accordingly.

The 11 Labs experience with media and entertainment illustrates exactly why speed across markets matters. The original thesis was sound — big studios generate massive content volume. Reality was that an industry heavily influenced by public perception and quality standards wasn't ready for the technology. Months were spent before the pivot to media creation platforms, which worked. Then the team spotted the agentic systems wave early, partnered ahead of it, and that bet is now enormous. None of those pivots were possible under a sequential "go deep then expand" model.

Corporate VCs are a stealth distribution weapon — get them on your cap table with contractual revenue commitments or don't bother

11 Labs has Woven Capital (Toyota), Deutsche Telekom, Telefonica, and Liberty Global on its cap table. None of them got there without strings attached.

The structure Reina negotiated is elegantly aligned: for every million dollars a CVC wants to invest, they must bring a contractually committed amount of revenue within 12 to 24 months. If they don't deliver, 11 Labs buys them out. This converts passive financial investors into active internal champions — people with every incentive to navigate procurement, route deals, and make the partnership work.

"Companies that have a CVC... they help you navigate big brands. They are your champions internally and their incentive is to actually make it work really well for the business."

The secondary benefit compounds over time: 11 Labs learned the entire telco industry by working closely with KPN, Deutsche Telekom, and Telefonica. They're learning automotive through Toyota's Woven Capital. Industry depth that would take years of enterprise sales cycles to accumulate is instead transferred through cap table relationships. When raising, the playbook is clear — target CVCs in your key verticals, negotiate pipeline commitments as a condition of allocation, and include buyout clauses as enforcement.

Customer Success is a revenue function — treating it as satisfaction work is ceding your expansion to a competitor who can spin up in two days

Chris Dagnon built Snowflake from zero to $4 billion telling people CS is complete bullshit — use professional services, charge for everything. Reina's counter is that Snowflake grew in a world where switching costs were real and competitors took years to emerge. That world is gone.

"Anyone can spin up a competitor of your product in the next two days. Anyone."

In that environment, CS that functions purely as a services business — charging for every interaction, treating the customer as a transaction — destroys the relationship depth that generates expansion revenue. The customers who stay and grow are the ones who feel like they're part of a community, not a billing cycle. CS needs to go as deep as possible, as early as possible, and its compensation should reflect that: quota on expansion, commission on upsells, not NPS scores.

The hunter-farmer tension is real but manageable. A hunter without tight coordination across a large account will create pricing discrepancies across subsidiaries that come back to haunt renewal conversations. The answer isn't to choose between hunters and CS — it's to run them together, with CS structurally positioned as a revenue generator from day one.

Brand is the most powerful sales cycle compressor in enterprise — three AI companies have already won the 'no one gets fired for buying IBM' position

Does brand dramatically reduce enterprise sales cycles? "Yes, one million percent. And whoever tells you no is just lying."

The target state Reina describes is explicit: become the asset that no procurement team gets fired for purchasing. The IBM of your category. In AI right now, exactly three companies have reached that status — OpenAI, Anthropic, and Cursor. Every subsequent sales motion those three companies run is cheaper and faster than any competitor's, because procurement teams default to them under pressure.

11 Labs is consciously building toward that position. The Formula 1 sponsorship with Audi Sauber (Revolut's team) isn't a vanity play — it's a calculated bet that a challenger brand building from scratch, competing against incumbents with technological conviction, maps directly onto 11 Labs' own narrative. The 11 Labs Summit in London was designed to prove a European company can run a flagship event at that scale. Executive dinners in key cities consistently outperform conference booths on ROI. The through-line: every brand investment compounds into every future enterprise deal.

Vertical sales teams kill performance when you implement them before pipeline has enough liquidity to sustain weekly deal closes

In India, Reina verticalized too early. The result: revenues depressed for an entire quarter, reps lost confidence, and the team had to go back to zero — full horizontal coverage, starting over from scratch.

The failure mode was threefold: reps weren't passionate enough about their assigned vertical, enterprise deal cycles in that segment were too long, and the team was too small for the volume required. When you segment into verticals before the pipeline can support weekly deal velocity, reps stop closing regularly, confidence collapses, and the entire market stalls.

His replacement framework is "pipeline construction" — a weekly calendar placeholder where he designs the ideal pipeline balance for each market. Every account executive needs both: big strategic whales that justify their enterprise-level comp, and enough mid-market liquidity to close deals frequently enough to stay motivated and confident. The signal to verticalize is when pipeline data shows clear enough volume per segment that each rep can maintain that balance independently. Until then, horizontal coverage and named account lists built around portfolio construction logic. The lesson from India cost a quarter — but the recovery was immediate once the structure was fixed.

The window is 18 months — and the companies that move now on brand, CVC partnerships, and parallel market entry will be impossible to displace

The broader signal from this conversation is about timing. CISOs and procurement teams are in an 18-to-24 month window of mandatory AI adoption — they have to have a message, have to have a deployment. That creates a brief period where even imperfect sales motions work. It won't last.

The companies that use this window to lock in brand recognition, CVC relationships with contractual pipeline commitments, and multi-market presence simultaneously will compound those advantages long after the window closes. The ones still running AI SDR blasts and sequential market entry will have trained their prospects to ignore them and ceded markets to faster movers.

The sales machine of the future is smaller, more elite, more AI-augmented — and structurally impossible to build if you wait.

---

Topics: AI sales, CRO strategy, sales compensation, quota design, go-to-market, customer success, corporate venture capital, brand building, enterprise sales, 11 Labs, outbound sales, sales team scaling