Fleet AI is moving into work that fleet software already automates
In recent announcements, Samsara, Motive and Fleetworthy have offered fleets AI tools that take on more of the administration around vehicles. Samsara made Agent Studio available to all customers in October, Motive launched Agents for safety and maintenance work, and Fleetworthy introduced RADAR inside its toll-management service in November. Across these releases, the ambition is to put software in charge of more of the work that follows a question or warning.
These launches enter a market already offering AI in many forms. AI capabilities feature across 243 supplier ranges serving the fleet and connected-vehicle market, including 160 with safety or incident-footage analysis and 93 with conversational fleet assistants, PAVE Insight has found. The competition therefore extends well beyond the suppliers making the latest announcements. An AI label gives a buyer little indication of which job the product can do.
But how much of the work now being promised actually needs AI? Maintenance work orders and configurable automation already coexist within the same product across 43 supplier ranges. A fleet assessing automatic job creation has a credible existing alternative to compare with the new agent offers.
Fleetio, for example, can turn an incoming fault into an issue, assign its priority and owner, then use the issue to create a work order. Its rules can also change vehicle status and notify staff. For an operator using that process, the new purchase needs to remove additional work or make the existing process easier to run.
There is a credible opportunity when that work needs interpretation or a conversation. Samsara describes voice agents that call drivers approaching a site, give instructions from company policies and answer questions. A geofence-triggered message can supply an entrance number. A conversation can handle a follow-up question the message did not anticipate, potentially reducing the calls and searches needed to obtain an answer.
That still leaves the question of what happens after the answer. AI that can carry out operational actions appears across 34 supplier ranges, while 24 suppliers have conversation and AI actions within the same product. The actions covered vary and can require human approval. Explaining an overdue repair, creating its work order and assigning the job remain separate capabilities to check.
The latest announcements show why those checks matter. Motive’s October explanation says customers currently choose triggers and actions, while creating workflows from ordinary language and setting goals for Agents to pursue are planned for early 2027. Fleetworthy’s early-access RADAR answers questions about toll activity and recommends next steps, with automated actions on its development plan. Both sit within the wider push into fleet administration, but the work they take over today differs.
The connection into the fleet’s other software can impose a further limit. Fleetio’s connection for external AI assistants reads and summarises account records, while its separate developer interface supports creating and updating work orders. An assistant connected through the first route can help someone understand outstanding repairs. Creating the repair job requires a supported route for changing the records.
So why are telematics suppliers trying to bring more of that work inside their own platforms? Well, the maintenance record, driver contact and operational data give them several parts of the job already. Motive describes Agents opening work orders in Motive Maintenance, while Fleetworthy is adding AI inside the toll service its customers use. Extending those relationships could bring more fleet administration onto the supplier’s software and strengthen its claim on the customer’s continuing spend.
For the fleet, that expansion earns its place when it improves the whole process. An operator already running a workshop system may gain more from connecting the new service to it than from moving maintenance records into another platform. The costs include setting up the connection, deciding who approves actions and keeping the two systems consistent. A faster first response can lose its advantage if someone then has to reconcile duplicate jobs.
There are encouraging results to test against that full process. Samsara reports that S. Bertram Foods cut weekly manual review time by 84%, using an agent to flag potential hours-of-service problems three hours ahead. That gives buyers a specific task to compare. Their own evaluation should include the time needed to check and act on the warnings, so the reduction in reporting work survives the rest of the process.
With existing automation and competing AI offers to compare, fleets can ask suppliers to demonstrate the additional work their product completes in the systems the fleet actually uses. A conversation may save a search or a phone call; a supported action may remove the next administrative step. The purchase earns its place through the work it takes off someone’s desk.
The workshop planner has a fairly practical way to judge the answer. When the agent has finished talking, is the right job ready for the right mechanic, or is the planner still the person who has to make it happen?
PAVE Insight publishes original connected mobility research and helps organisations apply the evidence to commercial decisions. Our work combines maintained market data, company and product analysis, and analyst interpretation.
If this analysis relates to a market, product, competitor or investment question your team is facing, discuss a research or consulting project. You can also explore our data or read more PAVE Insight research. Research and consulting can be commissioned without a platform subscription.