There’s an incredible amount of hype about the game-changing power of artificial intelligence (AI), but many experts agree the key ingredient to making the most of emerging technology is one thing — finding the right business use case.
Thierry Martin, senior manager for data and analytics strategy at Toyota Motors Europe, explains in a one-to-one video interview with ZDNET how the automotive giant is dedicating time and resources in research and development to the potential of AI.
Also: Data is the missing piece of the AI puzzle. Here’s how to fill the gap
However, this exploratory work is very much focused on the current use case — and that means data science rather than prediction and automation.
“The analysis of data is much more important for us,” says Martin. “For instance, how are people driving our cars? Is there a difference between different countries or highway driving between Germany and Belgium?”
The development of deep insight through data science and analysis is dependent on the collection of data, which is an area where Toyota excels.
“We can already get a lot of insight into how people are using our cars,” he says. “We do forecast models, for instance, to do root cause analysis or to predict what kind of accessories we need to install to help with planning.”
For now, Martin says Toyota is focused on using tools like Power BI to keep the human at the heart of the loop and to use analytics to develop a detailed understanding of automotive operations and processes.
“We are not letting the AI make decisions instead of people,” he says. “We prefer to provide more insight — that’s where we are.”
Yet in the not-so-distant future, Martin can envisage a situation where his organization starts to exploit AI in production — and explorations to find the right use cases for line-of-business processes are already underway.
“We have quite a high demand for that,” he says. “There are lots of use cases around analyzing text data and generative AI, which became possible since 2022 and the launch of the ChatGPT models.”
Also: Business success and growth is dependent upon trust, data, and AI
While OpenAI’s large language models (LLMs) helped push generative AI into the mainstream, Toyota — like so many other blue-chip enterprises — is proceeding with care when it comes to deploying emerging technologies.
Take the example of Omer Grossman, global CIO at CyberArk, who says his firm’s work around AI follows guidelines for safe and secure working that can be adopted and adapted.
“If you need a one-sentence slogan, this is it: Make sure you build responsible guardrails that promote innovation while keeping it secure,” he says.
In the case of Toyota Europe, Martin suggests two routes forward for making the most of AI.
The first pathway will focus on using tools like Microsoft Copilot at a personal level to help people complete tasks using non-sensitive data.
Also: What are Microsoft’s different Copilots? Here’s what they are and how you can use them
The second pathway, where his team is exploring its options through prototyping, is about using generative AI securely within the enterprise firewall to boost productivity.
“In terms of prototyping, we do a lot of work around chatbots,” he says. “We are coding chatbots ourselves now. And once you have a library set up, it’s very quick to set it up and try it by yourself. There’s not so much complexity here.”
Toyota Europe’s work with AI is being supported by the creation of a data mesh, which Martin describes as an approach to governance that ensures responsibility for data products stays with the business owners.
Also: Every AI project begins as a data project, but it’s a long, winding road
The organization is bringing its information together on a Snowflake platform that provides a foundation for well-governed data access.
The data mesh draws on a range of other technologies, including Dataiku for collaboration, Collibra for governance, and Denodo to connect data meshes across different parts of the organization, such as Toyota Europe and Japan.
Martin and his team are using these data mesh technologies to help explore AI. They’ve already built chatbots on Dataiku, which uses an LLM that runs on a secure instance of Azure Open AI to provide summaries of PDFs.
He’s demonstrated the chatbot to top executives at Toyota Europe and suggests that internal development is the way to go because it helps alleviate some of the concerns associated with publicly available models from big-name providers.
Also: The best AI chatbots of 2024: ChatGPT and alternatives
“So, we already have access to our own language model,” he says. “It’s on Azure, but it’s safe. And then on top of that, because we have the LLM and we have a chatbot, we can build our database and we can build interactive chatbots and things like that.”
Martin says his team continues to explore its options: “We are building a knowledge-retrieval system, for instance, because there is a lot of knowledge scattered everywhere in the company. But that’s still at the pilot level.”
Across the organization’s AI-enabled explorations, the watchword is “testing” to ensure that services meet tight governance requirements and the demands of line-of-business users.
“We want to confirm the value and we also need to confirm how to scale it,” he says. “Once you start to use a chatbot service, for example, there is AI ethics and governance that we have to bring in. If I start to roll out a chatbot, then I need to be clear on lots of questions about ethics.”
Also: Agile Intelligence: AI gives tech and business collaboration a much-needed boost
So, when might that broader implementation take place? Martin says he’d like to get some AI-based tools in production relatively quickly.
He’s working with his technology partners, including Snowflake, to ensure governance issues are considered and access to data is constrained.
“The vision should be, for instance, that a logistics chatbot should have access to only logistics data and not to HR data, just as an employee only has access to certain data,” he says.
Martin says Toyota Europe is continuing to prototype and could have some kind of AI-enabled chatbot service that extracts data from the Snowflake platform by mid-2024.
Also: How tech professionals can survive and thrive at work in the time of AI
He’s also speaking with other technology partners, such as Dataiku and Collibra, about how his vision might be realized.
Most crucially of all, he’ll be working closely with the business to demonstrate his AI services and to consider how these tools might work in specific areas of the organization.
“We need to understand where the best place is to run the chatbot,” he says. “And that’s why it’s super-important for the engineers and also the leaders to really understand what we are talking about. That’s where we’ll be spending our time.”