by
- Cliff Saran, Managing Editor
Learn how bet365 uses generative AI technology to understand legacy code in order to boost its modernisation program.
Alan Reed, head platform innovation at Hillside Technology, bet365, says that he’s never seen a technology spread as quickly and as widely as generative AI.
Reed describes GenAI as “transformational in a big T”
and says that GenAI technology was introduced to the mainstream two years ago with GPT 4. “Every time generative AI is mentioned, the word journey is used and we are no different,” says Reed. “We’re trying to understand it.” Reed says, “We’re trying to grasp its capabilities and our place in generative AI.”
Early adapters are eager to understand how GenAI can be used in their day-to-day tasks. This, he says is anything from an AI-based assistant to a tool which changes the way that people search for information, to using AI to do the heavy lifting in many organisations. He says that bet365 follows the same path. “We have a sliding-scale of ambition but obviously, like anything else we do in an organization of this size, we must measure, we must understand and we need to be very clear about what we are using generative AI for.” Reed points out that the entire tech sector is working to figure out how to make the most of the technology. GenAI can be used to write essays, but English is used instead of a programming language. But bet365 wants more than AI-based programming. He says: “Writing codes is great, but we are very curious to see if it can actually read code.”
Bet365 is particularly interested in using AI to understand code bases without having to read them. He says: “That’s appealing to us.”
“Like any large tech company or organization that has been around for many years, you begin to think about your technical data and legacy code base. You also start to consider that a portion of your workforce is dedicated to maintaining this legacy codebase.”
That was the original use-case for the team. “Most of the projects you implement begin to age and deteriorate the moment the code is live. “It’s more a maintenance model, and it becomes more complex,” he says. Reed says that if someone is not in the room when the code is being maintained and does not have access to the code, it may be harder to understand the code’s actual workings. “Every so often, you realize that you don’t really understand the code as well as you should for the task at hand,” he says.
Initially, the goal was to better understand the maintenance codebase of the business. Reed says that, like many other businesses, this code base has the potential to be modernised. “No organisation, no matter what size, is not trying to modernise their own code base.”
For Reed, AI must be a part of this. “We’re trying understand what that means for our business.”
Reed says that once AI provides a level understanding of the legacy code base it is possible to ask more complicated architectural questions. These questions are often raised by organisations as they modernise their IT system. This allows IT decision makers identify where functionality is duplicating and how to segment code to move an application to a cloud-native architecture. GenAI’s capability to read and understand code can help software developers identify code that can be replaced with better functionality or more performant code available in a repository.
