Aliisa Rosenthal

Aliisa Rosenthal is a general partner at Acrew Capital, investing in AI enterprise software and helping founders turn technical capabilities into businesses customers can adopt. OpenAI’s first commercial hire and former head of sales, she led the ChatGPT Enterprise launch and helped build the commercial organization as a self-service API business expanded into enterprise products.

From enterprise sales to AI investing

Rosenthal’s operating career spans several different stages of enterprise software. She held early leadership roles at Mixpanel and InVision, where she built the enterprise sales team. At WalkMe, she served as vice president of sales through its 2021 IPO, with responsibilities encompassing sales, account management, and renewals. That work extended beyond winning a customer’s initial commitment to managing the relationship after purchase and retaining the business.

At OpenAI, she faced a different commercial challenge: bringing rapidly developing AI capabilities to organizations ranging from startups to large enterprises. She helped expand its commercial organization from a small team to hundreds of people while launching multiple products. Enterprise adoption required more than access to a capable model; customers also needed security features, purchasing processes, suitable pricing, and help connecting the technology to their work.

Rosenthal joined Acrew in January 2026, moving from building commercial organizations herself to investing in companies and supporting their founders. She applies that operating experience to pricing, sales strategy, growth teams, and commercial hiring. Her perspective is especially useful where strong interest in AI has yet to become a repeatable business: deciding which customers to pursue, what they need before they will buy, and which demands justify a young company’s limited resources.

What commercialization taught her

ChatGPT made the gap between individual use and organizational adoption especially visible. After its release at the end of 2022, companies approached Rosenthal because employees were already using it, but their organizations needed single sign-on, nondisclosure agreements, and invoicing. The largest companies were the most vocal, steering the initial enterprise offering toward a high-end, expensive product. When OpenAI subsequently introduced a self-service option in January 2024, it grew faster and drew customers away from the sales-led offering. Many buyers preferred a cheaper product they could begin using without speaking to a salesperson.

That reversal informs her approach to building an AI commercial business: let actual adoption reveal where a more elaborate product or a human relationship is necessary. A large contract can be attractive while concealing substantial costs. Moving upmarket too early can pull a startup’s legal, security, engineering, product, and sales resources toward one customer before the company has established a repeatable business.

Automation must also preserve customer relationships. Rosenthal recalls herself and four OpenAI sales representatives overwhelmed by inbound demand. Prospective customers who received no acknowledgment sometimes bought elsewhere before the enterprise product was ready. Her lesson is practical: collect useful optional contact information, acknowledge inquiries, explain what is available, and ask what customers need. Her team used Clay to help manage demand; automated research and follow-up can free sellers to concentrate on qualified opportunities and substantive conversations.

Pricing creates another adoption barrier. Rosenthal describes the initial ChatGPT Enterprise price of $60 per user per month as an expensive lesson: organizations restricted purchases to selected teams rather than adopting it broadly. In her retrospective account, lowering the initial license commitment and shifting more of the charge toward usage brought in more contracts and allowed use to expand over time. Usage pricing introduces its own concern—unpredictable bills—so she recommends visible spending dashboards and optional caps. A manageable starting commitment lets customers discover value, while spending controls make expansion easier to approve.

Distinctive principles

• Start with self-service and automation. Rosenthal recommends building the automated commercial process before hiring a large team. Customers’ use of the product should reveal which capabilities and assistance they will pay more for; research, qualification, outreach, and routine administration can then run with less manual work. Add people where the process demonstrably stalls or buyers need trust, judgment, and a substantive conversation.

• Make the buyer’s next step easier. Sending a prospect away to complete an integration or gather internal approvals can leave a sale stalled indefinitely. Guided setup, live working sessions, or an on-site hackathon help customers make progress while the seller remains involved. Security reviews deserve the same attention: a trust portal can provide NDA signing and security documentation, while automated questionnaires reduce repetitive work. Human conversations should address the questions those resources leave unanswered.

• Separate self-service trials from resource-intensive pilots. A customer trying a product through a consumption limit, time limit, or restricted feature set can enter a scalable adoption funnel. A supervised proof of concept requires staff, ongoing support, and often another sales process after the experiment ends. Rosenthal reserves that work for particularly valuable opportunities. Customer references, demonstrations or evaluations using a prospect’s data, and contracts with opt-out provisions can sometimes answer purchasing concerns without committing a team to an extended pilot.

• Initial design partners need close collaboration. Her automation advice applies after the earliest customer-learning stage. The first partners need realistic expectations, frequent feedback, and a willingness to experiment with an unfinished product. She favors carefully chosen relationships, often through existing contacts or investor introductions. Beginning with an internal application at an established company can provide room to learn before taking on the demands of its principal production system, while opening a path to a larger agreement.

• Human trust becomes more valuable as routine work is automated. Rosenthal calls renewed demand for in-person relationships the revenge of the steak dinner. When outreach, demonstrations, and security paperwork become automated, meeting a person still helps buyers build confidence and connect a product to their own goals. She wants sellers who understand and use AI themselves, with enough judgment to explain the value of the change they are asking a customer to make.

Rosenthal also recognizes the tradeoff in hands-on implementation. Forward-deployed engineers are expensive and difficult to hire, but adapting a product to a customer’s workflows can make it more useful and harder to replace. The question is whether that investment creates a lasting customer relationship and enough value to justify the work.

Her experience at OpenAI similarly informs a qualified view of sales compensation. The company grew its sales team without commission plans during her tenure, but equity packages and the company’s increasing value helped make that possible. Early hires may be motivated by building the company and sharing in its upside; recruiting experienced enterprise sellers generally requires variable compensation. She favors simple plans that reward strong performance. At Acrew, these are concrete choices she helps founders navigate as they move from early customer learning to a commercial organization that can grow.