OpenAI Pays Zero Commission
Ross Sylvester, Co-Founder & CEO, Adrata | Feb 2026 | ~7 min read
The most valuable AI company on earth sells its enterprise product with zero sales commission. No accelerators, no spiffs, no President's Club. The company that could automate sales chose to build a 500-person human sales team -- and pay them like product managers.
I have been turning this over in my head for weeks. It breaks every assumption about how enterprise software gets sold.
The Playbook That Shouldn't Exist
Maggie Hott joined OpenAI when the entire go-to-market organization was fewer than 10 people. No SDRs. No SCs. No CSMs. No sales ops, no RevOps, no marketing enablement. Not even a working Salesforce instance. Two years later, she had scaled ChatGPT Enterprise's sales team to 500 people, built during a period when OpenAI went from 250 to nearly 1,000 employees company-wide.
The growth is staggering. But the compensation model is the real story.
Zero commission. Every seller at OpenAI earns a fixed salary. No variable comp tied to bookings. No accelerators after 100% attainment. No quarterly clawbacks. The enterprise AE OTE averages around $290K-$305K, with an 82/18 base-to-variable split -- but the variable isn't traditional commission. It's tied more broadly to team and company outcomes, structured closer to how a product manager or engineer gets compensated than how a traditional AE does.
Top performers can earn between $310K and $900K. But the path to the high end isn't "crush quota and collect accelerators." It's "be the person who figures out how to sell a product that didn't exist two years ago."
This is not entirely unprecedented. Slack famously started with no outbound sales team at all -- 100% inbound, land-and-expand. Stripe built its billion-dollar sales organization with a similar philosophy of hiring for problem-solving over process execution, where reps manage the entire sales process from start to finish -- they source, engage, qualify, demo, and sign contracts with no awkward handoffs. But OpenAI is doing it at a scale and velocity that neither Slack nor Stripe attempted in their enterprise infancy.
Think about what this means structurally. The typical enterprise sales cycle at OpenAI runs approximately 55 days from first conversation to closed-won, with an estimated annual quota per enterprise AE of roughly $1.45M. That's a 5x OTE-to-quota ratio, which is standard. But the mechanics of how reps hit that number are completely different. There's no "end of quarter push." No "I need three more meetings this week or I miss accelerators." The pressure is to do excellent work for the account, not to close by the 30th.
The "Chaos Translator" Thesis
Here is the part that I find most interesting. When you remove commission from the equation, you fundamentally change who you hire.
Traditional enterprise sales hiring optimizes for closers. People who can run a process, manage a forecast, hit a number. The commission structure self-selects for that profile. The top earners are the people who are best at the mechanics of closing deals within a defined playbook.
OpenAI doesn't have a defined playbook. When Hott was building the team, everyone had been at the company less than six months. New roles, new managers, new workflows -- constantly. The product was evolving weekly. The use cases were being invented by customers in real time.
So OpenAI hired for a completely different profile: ambiguity tolerance. The ability to function -- and sell -- in an environment where the product roadmap, the pricing model, the competitive landscape, and the org chart are all moving simultaneously. I call this the "chaos translator" archetype: people who can take organizational uncertainty and translate it into customer confidence.
The result? OpenAI claims a 100% pilot win rate. Every customer who runs a pilot converts. That is an extraordinary number. It suggests that the bottleneck isn't closing -- it's getting the right accounts into pilots in the first place. And when that's your bottleneck, commission on individual deals is irrelevant. You need people who can identify the right accounts, navigate ambiguity, and let the product do the closing.
The Stack Behind the Sellers
OpenAI's sales team doesn't just use AI products. They use AI to sell AI. The internal tooling is as interesting as the compensation model.
Clay for targeting. OpenAI uses Clay for lead enrichment, account research, and targeting. Their RevOps team doubled enrichment coverage from the low 40s to the high 80s through Clay's implementation, running 8,500+ total enrichments via Clay's on-demand Salesforce package. They used Clay's AI research agent to mimic what their best sellers did manually -- visiting company websites, scanning LinkedIn, looking for revenue figures, recent developments, and significant changes in the last 90 days -- and then automated it at scale.
Internal GTM Assistant for meeting prep. OpenAI built a Slack-based tool that generates daily meeting briefs with account history, call notes, Salesforce activity, and product release updates. Teams average 22 messages per week with the assistant across briefs, recaps, and Q&A. Users report a 20% productivity lift -- roughly one extra day per week freed up to spend with customers. The tool is described internally as "like having a virtual coworker that we re-skill every single week."
The GTM Assistant is now piloting capabilities to log CRM updates after calls automatically, spot noteworthy usage patterns, and draft customer follow-ups -- all running in the background without rep involvement.
This is the part that should make every CRO pay attention. OpenAI isn't just selling AI. It's building the operational infrastructure to sell with AI, and that infrastructure is what enables the zero-commission model. When your tooling handles the research, the prep, and the follow-up, you can afford to pay sellers for judgment rather than hustle.
The Broader Signal
OpenAI's model is not happening in isolation. Across the top AI companies, I see a consistent pattern:
Enterprise is the priority. Roughly a third of all open GTM roles at top AI companies target enterprise accounts. These aren't PLG-only businesses anymore. They're building serious enterprise sales motions.
Anthropic is building outbound. Anthropic is actively hiring a Manager of Sales Development to "build, lead, and scale a team of 8-12 BDRs focused on both inbound lead management and strategic outbound prospecting." The compensation range: $165K-$245K base. That's a real investment in a real outbound motion. Even the company that makes Claude -- which could theoretically automate outbound entirely -- is hiring humans to do it.
Compensation is converging. Enterprise AE OTE at the top AI companies averages $300K-$320K. That's competitive with top-tier SaaS companies but not the $400K-$500K OTEs you see at enterprise infrastructure companies. The comp is good. But it's not "close one mega-deal and make your year" money. It's "be a great operator across a portfolio of accounts" money.
The common thread: these companies are building sales organizations that look more like customer success organizations with revenue responsibility. The sellers are experts. The tooling handles the process. And the compensation rewards consistency over heroics.
There is an irony here that I keep coming back to. The companies building the AI that will automate sales are choosing to build human sales teams. Not small, scrappy teams -- 500-person organizations with dedicated SDRs, SCs, CSMs, and RevOps. If anyone could prove that AI replaces human sellers, it's OpenAI and Anthropic. Instead, they're proving that even the most advanced AI products still need humans to navigate enterprise complexity. The humans just look different than they used to.
What I'm Watching
Does zero-commission survive contact with competition? Right now, OpenAI has extraordinary demand pull. Everyone wants ChatGPT Enterprise. The pilots sell themselves. But what happens when enterprise AI becomes a competitive market -- when Anthropic, Google, and Microsoft are all pitching the same CISO? Does the model hold when deals require real competitive selling, objection handling, and political navigation? Or does OpenAI quietly introduce commission when the easy growth phase ends?
Will this model spread? If zero-commission works at OpenAI's scale, why wouldn't every high-growth PLG company adopt it? The argument for commission has always been that it aligns seller incentives with company outcomes. But if the product sells itself and the real job is account selection and customer enablement, maybe fixed comp creates better alignment than variable comp. That's a heretical idea in enterprise sales. But OpenAI is generating data to support it.
What happens to the sales profession? The top AI companies are hiring sellers who look like product managers, paying them like engineers, and equipping them with AI tooling that automates the traditional sales workflow. If this is the future, the skill set that defines a great seller in 2030 looks nothing like the skill set that defined one in 2020. The quota-carrying closer may be replaced not by AI, but by a human with a completely different job description.
Sources
- How We Scaled OpenAI's Sales Team from 10 to 500 People in 2 Years - SaaStr
- OpenAI's GTM Lead Maggie Hott: No Commissions, 100% Pilot Win Rates - SaaStr
- How OpenAI is Scaling Their GTM Motion with Clay
- Driving Sales Productivity and Customer Success at OpenAI
- OpenAI Enterprise Account Executive Salary - RepVue
- Manager, Sales Development - Anthropic
- How We Built ChatGPT Enterprise's Sales Team from Absolute Zero - SaaStr