For nearly 100 years, Procter & Gamble has been known for innovating products that become indispensable to consumers around the world, from dish soap to diapers to denture adhesive.
Now, it’s on the frontlines of another massive transformation because of the rise of artificial intelligence. On a recent episode of Where AI Works, Wharton professor Christian Terwiesch spoke with Alfredo Colas, who is chief data and AI officer at Procter & Gamble, about how the company is reinventing itself again. With nearly three decades at P&G, Colas shared his perspective on embedding AI into every aspect of operations while keeping humans firmly in the loop. He discussed how the company leverages AI to supercharge creativity — turning individual ideas into high-performing concepts — and how AI helps democratize knowledge, making decades of consumer insights instantly accessible to employees across the company. Their conversation also highlighted the importance of human judgment in final decision-making, the evolving nature of work, and how AI can eliminate routine tasks to let employees focus on creative problem-solving.
Where AI Works is produced by Wharton in conjunction with Accenture. Listen to the episode, or keep reading for key highlights from the conversation.
AI Is a ‘Cybernetic Teammate’
Procter & Gamble is capitalizing on one of the most effective capabilities of AI: helping employees generate a lot of ideas very quickly. During hackathons at Wharton and Harvard, the company divided more than 700 employees into four groups that worked either in teams or as individuals, with or without AI.
They found that an individual with AI performed better than a team without AI. And the team with AI generated most of the ideas that were assessed by professors and business leaders to be among the best.
“What we saw is this new concept of the cybernetic teammate, where we could see AI helping both our marketing and our R&D specialists come up with better and more holistic ideas than when they were working individually or even when they were working together, but without the support of AI,” Colas said.
“The moment we removed all that friction, made it very easily accessible to everyone, consumption skyrocketed.”— Alfredo Colas
AI Can Unlock Hidden Demand for Knowledge
AI is dramatically increasing access to information by removing friction, which is helping to drive productivity and quality at P&G. Colas gave a powerful example of how AI democratizes knowledge: A central team used AI to develop a database that made much of the knowledge in one domain available to everyone in the company. When the database launched, the system got more queries in the first month than the central team had received in the last 10 years. Colas said the numbers prove that there was always a desire to access that knowledge, but perhaps employees didn’t want to ask someone.
“The moment we removed all that friction, made it very easily accessible to everyone, consumption skyrocketed. And that was knowledge that has been part of P&G all along,” he said.
Consumer Insight Still Starts With Human Observation
Colas said P&G pioneered the concept of customer data by sending representatives into people’s homes to ask them how they used the company’s products and their likes and dislikes. That first version of “big data” was scribbled in notebooks. It was later automated by computers, and AI is now doing even more with it. But Colas said the technology isn’t “a complete replacement of the formula.”
AI can summarize, organize, and expand on data without bias, but it cannot provide empathy and nuance. Humans must remain in the loop, he said, because the strongest insights come from combining human intuition with AI’s analytical power.
“It does not replace that first-party data that goes into the consumer visit, talking to them and understanding their concerns, their daily lives, their habits, and from there gaining relevant insights about problems they are facing or opportunities to come up with products,” Colas said.
“AI is a means, not an end.”— Alfredo Colas
Human Judgment Still Drives the Final Decision
Terwiesch questioned Colas about the need to keep humans in the loop, especially when it comes to taking responsibility for decisions that involve the use of AI.
“In our research here at Wharton, we show that AI rocks at generating ideas. But at the end of the day, somebody has to make a decision,” he said.
Colas agreed. He emphasized that while AI can expand the number and quality of choices, company leaders must still approve the decision and are held accountable for the outcome. AI can support that decision-making, but that’s as far as it can go.
“As I always tell my team, you cannot come back and say the old student excuse that the dog ate my homework. You cannot come back and say the product was a flop because gen AI made that recommendation. No, a human signed that recommendation and a human approved it,” he said.
What’s the Next Frontier for AI?
Colas paraphrased former Canadian Prime Minister Justin Trudeau, who told the World Economic Forum in Davos that the pace of change has never been this fast, yet will never be this slow again. It’s a stark reminder of the speed at which technology moves. Colas said he prefers to think positively about the future of AI, focusing on how it will take more toil out of the everyday so that humans can elevate their work, and how it will help personalize communication for different audiences.
He offered advice for anyone in a setting where AI is deployed: Use AI in meaningful ways that create value for the company; and invest in employees because they are the ones who will make AI succeed or fail.
“AI is a means, not an end. So be clear on what end you’re trying to achieve,” Colas said. “What is the business objective that you’re trying to achieve, and make sure you have the leadership of your company aligned and supportive of that direction.”
This article was partially generated by AI and edited (with additional writing) by Knowledge at Wharton staff. Read our AI policy here.
