Increasing use of AI in fleet management is shifting the emphasis in the sector from data to insight, FleetCheck is reporting.
Peter Golding, CEO of the fleet software specialist, said the rise of large language models (LLMs) such as Claude and ChatGPT meant industry professionals wanted to make direct requests rather than interrogating data by traditional means.
‘We have gone through a period of at least two decades when fleet managers have been deluged with an accumulating amount of data, with a growing emphasis on finding ways to mine that information in ways that create understanding.
‘Fleet software user interfaces have become ever more sophisticated as a result, with a range of approaches from the highly complex to our own design, which emphasises key metrics. These can be highly effective, but the new technology offers the opportunity for a rethink.’
Instead of using a dashboard to drill down into data and identify pertinent findings, LLMs allow natural language questions to be asked and meaningful results generated.
Peter said, ‘In a simple example, a fleet manager may today use a custom dashboard within their fleet software to list all their drivers falling short of WLTP miles per gallon figures by more than 25%. This is relatively easy, but not as easy as asking Claude.
‘As the insight required becomes deeper and the questions more technical, the use of AI should offer easier access to a wider range of insights, which are only really limited by the fleet manager’s curiosity.
‘In particular, there are likely to be benefits where different functions within a company – not just the fleet but perhaps finance and human resources, too – want to use the information available for different purposes.’
It was increasingly likely fleet managers would want to use the LLMs to which they were accustomed as the front end of their fleet software, with the system itself effectively becoming invisible in the background, he added.
‘This is one direction that AI could take for fleets that has obvious benefits, and we are examining ways to make this happen as seamlessly as possible,’ concluded Peter.