Aliisa Rosenthal advises AI startups to prioritize automated, self-serve sales models before building large sales teams, emphasizing simplicity, frictionless buying experiences, and strategic human support only when necessary. Drawing from OpenAI’s experience, she highlights launching self-serve products first, adopting flexible pricing, and employing AI-native salespeople to scale effectively while maintaining customer satisfaction.
Aliisa Rosenthal, who joined OpenAI over four years ago, shares her insights on building go-to-market strategies for AI products based on her experience scaling OpenAI’s sales organization from a small team to a multi-billion dollar enterprise revenue generator. She advises founders to reverse the traditional sales approach by starting with automation and self-serve tools before building a large sales team. This approach leverages the many modern AI-driven tools available in 2026 to automate inbound and outbound sales processes, allowing companies to identify bottlenecks and add human support only where necessary.
Rosenthal recounts OpenAI’s experience launching ChatGPT Enterprise, initially focusing on a high-end, expensive enterprise product based on feedback from large companies. However, when they introduced a self-serve version four months later, it rapidly outpaced the enterprise offering, revealing that many customers preferred a frictionless, no-salesperson-needed experience. This taught them to launch self-serve options first, learn from customer feedback, and then build more complex enterprise offerings with human sales support, rather than the other way around.
She emphasizes the importance of making the buying process as easy and frictionless as possible for customers. This includes minimizing the number of pricing options, simplifying paperwork, automating security processes through trust portals, and avoiding pilots or proofs of concept (POCs) unless absolutely necessary. Rosenthal suggests creative alternatives to pilots, such as demos with custom data or contracts with opt-out clauses, to maintain control over the sales cycle and reduce resource drain.
Regarding pricing, Rosenthal shares that OpenAI initially priced ChatGPT Enterprise too high, which limited adoption. They later shifted to a lower barrier-to-entry pricing model with a license fee plus usage-based charges, which significantly increased customer uptake and product usage. She also discusses when to hire salespeople, recommending waiting until there is clear demand for human interaction, especially for larger enterprise deals, and hiring AI-native salespeople who genuinely understand and believe in the product.
Finally, Rosenthal touches on the role of forward deployed engineers, who can make products stickier by providing customized integration support but are expensive and hard to hire. For startups working with their first customers, she advises a highly personalized, hands-on approach with carefully selected design partners rather than automated outbound tools. Overall, her key message is to build scalable, automated sales machines first, then layer in human support strategically to optimize growth and customer experience.