Fleet Managers are already using artificial intelligence to write policies, analyse spreadsheets, summarise documents and speed up routine administration.
The bigger shift, however, will not come from asking a chatbot better questions. It will come when AI can safely use an organisation’s own data to complete repeatable fleet tasks.
Eden Shirley, Managing Director and Founder at FleetGuru.ai, believes the industry is moving from personal productivity tools towards operational AI, but most organisations are not there yet.
“I still feel like we’re coming out of just the personal use benefit phase, and now moving into a corporate operational phase where everybody’s trying to get their data ducks in a row,” Shirley said.
Personal AI is already saving time
The current generation of AI tools is useful because it is accessible.
A Fleet Manager can ask a large language model to draft a position description, review a procedure, organise information or explain a spreadsheet. These tasks can deliver an immediate productivity improvement without requiring a major technology project.
Shirley described this as the personal benefit stage of AI.
The user is still responsible for checking the result, making decisions and moving the information between systems. AI helps complete the task, but it is not yet running the fleet process.
Enterprise AI is different because it must work consistently across the organisation.
“At an enterprise and operational level, you really need those solutions to be repeatable and accurate,” Shirley said.
The fleet system needs to know the business
For AI to make operational decisions, it needs more than general knowledge. It needs access to trusted data about the organisation.
That could include vehicle locations, maintenance records, fuel use, work orders, sales forecasts, customer demand, operating costs and accounting information.
When those data sources are connected, the system can begin to identify relationships that would normally require several reports and spreadsheets.
For example, an organisation could combine expected sales activity with fleet locations and maintenance schedules to determine where vehicles will be needed over the next six months.
That information could influence vehicle allocation, servicing, replacement timing and purchasing decisions.
“Let’s get the accounting information out of the accounting software. Let’s get the sales information out of whatever the sales CRM is. Let’s get the fleet information out of the vehicles,” Shirley said.
“Then you’ve really enriched the data.”
Controlled answers will matter
One of the risks with general AI tools is that they can produce a confident answer without understanding the organisation’s rules or the technical requirements of a specific vehicle.
Enterprise fleet applications need a more controlled approach.
Shirley explained that FleetGuru.ai combines language models with structured maintenance information, including service and repair data published by vehicle manufacturers.
The language model interprets how the user has phrased the request, while the structured information determines the correct result.
“It is not an LLM making up what should happen to a car,” Shirley said.
This model allows users to communicate naturally while reducing the risk that the system invents a maintenance requirement or recommends an action that does not match the vehicle.
Fleet tasks will move into everyday tools
Operational AI may also change how Fleet Managers interact with fleet systems.
Instead of logging into several platforms, exporting reports and sending emails, fleet tasks could be completed from Microsoft Teams, Slack or another communication tool.
A Fleet Manager could receive a notification that several vehicles are due for servicing, approve the work in the same conversation and allow the system to arrange the bookings.
“You should actually be able to just be in your notification on a mobile phone,” Shirley said.
“If you got a notification that said these are due, you just turn around and type back, ‘Okay, get them scheduled’.”
This would move fleet management away from manually checking systems and towards managing exceptions, approving actions and reviewing outcomes.
Suppliers will deliver the early benefits
Building an organisation-wide AI capability will require investment in data quality, security, privacy and integration. For many fleets, that process will take time.
In the near term, Fleet Managers are more likely to access operational AI through their existing suppliers.
Telematics providers, fleet management organisations, maintenance platforms and software companies are already developing tools that can use fleet data in a controlled environment.
“The major service providers are obviously leading the charge,” Shirley said.
This means Fleet Managers should be asking suppliers practical questions about where their AI gets its information, how the output is checked, what data it can access and which tasks it can complete.
AI may already be improving personal productivity, but the larger fleet benefit will arrive when reliable data, trusted systems and automated actions are brought together.
That is the point where AI stops being another software feature and starts changing how the fleet is managed.
