ChatGPT’s ultimate use case in business will be knowledge management, says EY CTO

Currently, there is no “killer” use case for using ChatGPT in the enterprise – i.e. one that will have a huge impact on the revenue. business and bottom line – according to EY’s global chief technology officer, Nicola Morini Bianzino.

But that could soon change: The next six to 12 months will bring an explosion of experimentation, he predicted, especially once companies are able to build on ChatGPT using OpenAI’s API. . And the killer use case emerging could be about the impact of generative AI on knowledge management – ​​which Bianzino describes as the “dialectic of AI”.

“Knowledge companies tend to store knowledge in a very flat two-dimensional way, which makes it difficult to access, interact and dialogue,” he told VentureBeat in an interview. “We tried 20, 30, 40 years ago to build expert systems. It didn’t go very well because they were too stiff. I think this technology promises to overcome many problems experienced by expert systems. »

As ChatGPT and similar tools evolve and improve, and can be trained on a company’s data in a secure way, it will change the way we access and consume information within the company, he explained.

“I think we’ll get to a point where we can actually have a conversation about business performance with an AI agent,” he said. “You query the system, the system is able to maintain a state of the conversation, and then each question allows you to dig deeper into the problem and understand it better, instead of letting me write a report on sales in that particular region for the last month, which usually doesn’t provide much information.

A significant impact on enterprise software

This opportunity for the future of generative AI would have a significant impact on enterprise software, Bianzino explained, as organizations should start thinking about new ways to structure data within an enterprise that goes to the beyond traditional analysis tools.

“To be fair, there are nice dashboards, much better data structures, but not huge value,” he said of the current tools.

On the other hand, Bianzino suggested imagining a future where a ChatGPT or equivalent could be “invited” to a board meeting and be asked questions.

“The feature that impressed me so much is the tool’s ability to maintain conversation state,” he said. “With a regular, more conversational agent, you ask a question and get an answer. With ChatGPT, you can go deeper, you can say, tell me, what went wrong last week? And then you say, we didn’t meet our targets, okay, what products we didn’t sell? And then you can say, what about the weather in the area? So you can go down a parse tree that is not predetermined.

ChatGPT should overcome business hurdles

At this point, the opportunity described by Bianzino is moot. There are many different hurdles that ChatGPT and other generative AI tools have to overcome, from potential ethical implications to accuracy issues that even OpenAI CEO Sam Altman has admitted.

Text and documentation generation should also be trained and aligned with the appropriate ontology of the specific enterprise, Bianzino explained, and should be securely contained, stored, and controlled within the enterprise. But if it could be, he said, it would create an enormous amount of value.

“When you think of an organization like ours, we have 360,000 people, we have a lot of tools and capabilities built over the more than 100 years of our history,” he said. “But that knowledge is distributed now, you can’t really touch it, it’s the soul of our organization, but it’s intangible.” If you could systematize it into an ontology and integrate it into a technology solution, you can dramatically increase business value, he continued.

“Whenever I talk to mining customers or other types of resource customers on the right, they all lament the fact that there aren’t as many engineers from academia as there are. there were in the past,” he said. “They worry about losing the knowledge of people retiring – so consider storing that knowledge and making it accessible through a tool that gives them easy access to that value.”

Experimentation and an explosion of use cases are on the way

Bianzino says EY is currently working on the possibilities: “We’re putting a team there and experimenting where we can take it – we need to better understand how to structure the underlying data to extract the most value possible from it. interaction,” he said, adding that it’s not yet clear whether unstructured data such as videos or music can be processed. “But I’m very optimistic that we can deploy something. something like this in a short period of time for a company like ours.”

While still in its infancy in the generative AI landscape, Bianzino encourages organizations to play around with tools like ChatGPT, even just for fun.

“One of my kids showed me yesterday that you can create a story with ChatGPT,” Bianzino said. “If you start playing with it, I think you start to understand the potential.”

Business leaders, such as the CTO and CIO need to be aware of these trends, he continued, because unlike something like quantum computing, which could still be 10 to 15 years away, true generative AI opportunities may only be 6 -12 months away.

“It’s going to be big,” he said. “At the moment, there are no killer use cases…but I think it will go very quickly.”