The executives leading the product and engineering efforts for OpenAI’s developer platform say companies must adopt a result-oriented approach to successfully roll out AI to employees.
Olivier Godement and Sherwin Wu head product and engineering for OpenAI’s developer platform, respectively. During an interview on the BG2 podcast that aired Thursday, Godement and Wu shared three tips on how companies can integrate AI.
“Number one is the interesting combination of top-down buy-in and enabling a very clear group, like a tiger team,” Godement said. He added that the team could be a mix of staff from AI providers like OpenAI, as well as the company’s own employees.
Godement said members of the “tiger team” should possess either technical skills or a deep understanding of the company’s processes.
“In the enterprise, like customer support, what we found is that the vast majority of the knowledge is in people’s heads,” Godement said.
“Unless you have that tiger team, a mix of technical and subject matter experts, it’s really hard to get something out of the ground,” he added.
Next, Godement and Wu said companies need to develop clear benchmarks, or what they call “evals,” to track their progress with AI.
“Evals are much harder than what it looks to get done,” Godement said.
“Evals, oftentimes, need to come bottom-up, because all of these things are kind of in people’s heads, in the actual operator’s heads. It’s actually very hard to have a top-down mandate,” Wu said.
Lastly, Godement said that companies should monitor their benchmarks closely and strive to make progress against them.
“A lot of that is like art sometimes, more than science,” he said.
Related stories
Business Insider tells the innovative stories you want to know
Business Insider tells the innovative stories you want to know
Progress can be achieved by having a good understanding of an AI model’s design, behavior, and constraints, Godement said.
“Sometimes, we even need to fine-tune ourselves the models, when there are some clear limitations, and you know, being patient, getting you way up there and then, ship,” he added.
Godement said it was important for a company’s top leadership to make AI a priority and give their staff the opportunity to experiment.
“Letting the team organize and be like, ‘OK, if you want to start small, start small, and then you can scale it up.’ That would be number 1,” he added.
OpenAI did not respond to a request for comment from Business Insider.
Tech CEOs have been stepping up their efforts when it comes to getting their employees to use AI.
Duolingo CEO Luis von Ahn said in an August interview with The New York Times that Duolingo has been organizing weekly activities to encourage teams to use AI.
“Every Friday morning, we have this thing: It’s a bad acronym, f-r-A-I-days,” von Ahn told the Times.
Howie Liu, the CEO of the vibe coding platform Airtable, said in an episode of Lenny’s Podcats that aired last month that he wants his staff to experiment with AI, even if it entails taking time off work.
“If you want to cancel all your meetings for a day or for an entire week and just go play around with every AI product that you think could be relevant to Airtable, go do it. Period,” Liu said.