Budi Saputra, Lorenz Maierhofer, and Rafal Kawala suspected they had a success on their hands when 8,000 requests to try ChatPG came through in the first two minutes. Today, more than 30,000 users are already putting the company-wide GenAI underpinning ChatPG to work across different business models, from upskilling to onboarding.

The three Signal Innovators for 2024, Saputra, Maierhofer, and Kawala believe that by not locking themselves into one model, such as chat, they expanded how the tool could be adopted throughout the company and enhance multiple divisions.

“Our strategy to be flexible is a recipe for success, and we can firmly state that we’re ready for new models, for new tools, for new AI,” said Kawala.

You can hear more from this innovative team in the video below or read our lightly edited transcript.

TRANSCRIPT

Rafal Kawala
Our efforts around generative AI started around November 2022, of the release of ChatGPT, which created a massive buzz in the industry.

Lorenz Maierhofer
This is not only technology for the sake of technology, but it’s bringing this to real world business applications, and therefore bringing it to an impact.

Budi Saputra
How do we build our algorithms? How do we bring our data and processes and transform them to benefit P&G.

Maierhofer
Before this excitement, we had to find a way to manage the risks

Kawala
And filtering with what worked from what didn’t. Was the super challenging tasks.

Maierhofer
There is hallucination in these models. There is lack of accuracy metrics. There is issues with math,

Kawala
And we also had to deploy AI responsibly with all of the governance controls and security in place. So that was the start of the whole process. when

Maierhofer
When we made the decision that, yes, there is a strategy that requires a chatPG, askPG and imagePG, we very quickly stood up a better version that we started testing, not only with a handful of people, but with multiple 1000s of people. And that was instrumental to learn fast and validate where the value is, and led us to a company wide launch within about two three months.

Saputra
The first two hours, when we launched at PG, we received 8000 requests, which is really blow my mind. How much excitement and how much interest and how much engagement the employee has.

Kawala
We had to scramble around and communicate with the whole PNG, basically that was interested in the technology. So there was simply so much to organize and understand.

Saputra
And today, fast forward, if you will. We are more than 30,000 of users already.

Kawala
It was extreme amount of effort to understand what and how we approach it, and how we enable others to kind of start piloting these things. It doesn’t have reference architecture, reference evaluations, reference governance, and we have to figure out control, safety, make sure that P&G data assets are correct.

Saputra
We put together governance training as well that every single employee that wants to get access, they need to pass this training.

Maierhofer
Gen AI becomes one extra tool in our tool belt. It’s not that we are approaching every problem as a Gen AI problem, but of course, Gen AI opened the door for a lot of new discussions.

Saputra
People get into this technology with their eyes wide open, understanding the benefit, but also be aware of the potential of this as well.

Maierhofer
And we, from the beginning on, basically set the strategy of, let’s not lock ourselves into one specific choice, but build a Gen AI platform that allows us to kind of switch between different models, as we saw necessary from a business perspective.

Saputra
I know how much work you know, not only the product management, but the data scientists and the AI engineers have put in, the blood, tear and sweat that get into this.

Maierhofer
The individual teams have their own AI champions embedded. They run their own upskilling trainings. It’s becoming of the standard onboarding procedures. That’s the way to make this the new normal.

Kawala
Our strategy to be flexible is a recipe or success, and we can firmly state that we’re ready for new models, for new tools, for new AI.