OpenAI lead: Survival demands a 3-month AI product plan

This article summarizes an episode of Lenny’s Podcast’s video series featuring Tara Seshan, product lead at OpenAI.

Tara Seshan, product lead at OpenAI / Photo credit: Lenny’s Podcast

Building AI software based on what is possible today, or what might be possible in a year, can quickly make a product obsolete. Tara Seshan, a product lead at OpenAI, argues that market survival requires a strict two-to-three-month planning window.

She says that companies must treat AI agents as coworkers, use the technology for capability expansion rather than simple automation, and push development teams to aim higher.

AI agents require familiar workspaces

To turn these higher aims into useful products, AI needs to fit into daily workflows rather than operate in complex, isolated environments:

• Familiar interfaces: Start with simple chat interactions and handle model selection automatically.
• Background processing: Keep complex tasks running on remote servers so progress continues even if the user logs off.
• Secure access: Provide the AI with the same system permissions a human coworker would need.
• Human oversight: Allow human workers to review the tool’s output before granting it independence.

This approach makes AI a more integrated part of the team. Seshan predicts that the next era of enterprise software will involve employees treating AI as a coworker that can collaborate with multiple people toward shared goals.

Capability expansion eclipses basic automation

Treating AI as a collaborator shifts its value beyond routine task completion.

Successful professionals use these tools to take on new projects, as Seshan points out that “the people we see who are most effective at using AI tools don’t simply use AI to automate rote tasks, but use it to expand the set of things they are capable of doing.”

Executives must recognize this difference by focusing on capability expansions:

• Redefine daily work habits. Advanced features only yield results when employees alter how they approach projects.
• Delegate higher-level responsibilities. Workers generate the most value when they hand over strategic execution to the technology.
• Capture new market share. Building new capabilities creates a competitive advantage over rival firms stuck on cost reduction.

Short horizons demand empirical planning

Because AI capabilities change quickly, product teams need to build for what the technology can realistically do in the near term rather than betting on distant improvements.

Seshan warns that “you fail if you build for where the models are now. You fail if you build for where you think the models will be in a year. Both outcomes are equally wrong.”

To survive, development teams must organize their planning around testing assumptions and gathering customer feedback before the technology changes again: