Artificial intelligence is enabling startups to generate higher revenue with smaller teams while changing hiring, engineering and product roles, said Marc Manara, head of global Startups, Venture Capital and Private Equity teams at OpenAI, adding that AI models are expected to become significantly cheaper over the next year.
Speaking at FutureLab: Building Defensible AI Companies, organised by Newton School of Technology on Thursday, Manara said one of the biggest changes he has seen among startups is that teams are becoming smaller while using AI across different parts of the business.
"I met a team last week that's eight people doing tens of millions in revenue, and they're using a lot more AI to run the business," he said. "Yes, on the engineering side, but also how they do interviewing, hiring, how they're doing customer support... they're trying to build agents to offset what would have been a headcount."
Manara also said AI is changing the way software teams are organised, with traditional roles beginning to overlap. He said startups are hiring fewer product managers while engineers are taking on more product responsibilities.
"Most people agree that product managers and engineers as a role are blending, and engineers are doing product work more and more, and product managers are shipping code,” he said.
Manara added that many startups are also creating dedicated AI engineering teams focused on model evaluation and model selection, describing it as a specialised skill set.
Another role gaining importance is the forward deployment engineer, who works closely with enterprise customers to help implement AI systems. Manara said OpenAI has invested in such teams because many companies want to use AI but lack the expertise to apply it to their own workflows.
"It's probably one of the hottest job titles right now in Silicon Valley," he said.
Manara also said he expects the cost of using current AI models to continue falling over time, making advanced AI capabilities more affordable for developers.
"I think you can pretty safely bet that the cost of current intelligence, even the current frontier intelligence, will come down," he said. "In six or 12 months, that equivalent intelligence will be much, much cheaper."
Speaking about India, Manara said the country's technical talent gives founders an advantage and noted that many Indian startups are already building products for customers outside the country.
On where startups should compete, Manara said founders should not be concerned about competing directly with frontier AI labs. Instead, he said companies should focus on specialised markets.
He added that startups serving niche customers have an advantage because large AI companies have limited resources and cannot build products for every specialised use case.
"I think a startup that's laser focused and specialised and purpose built for that thing generally has a really good shot," he said.