The fleet manager of a 140-truck regional carrier does not open a trade directory when a maintenance contract comes up for renewal. She opens ChatGPT, types a question that includes her fleet size, her lanes, and the thing that went wrong with the last vendor, and reads a shortlist of four names. The call she makes next is to one of those four. Over the past two quarters we have logged how the major AI engines answer that kind of question across the four vendor categories fleets buy from most, and this piece lays out what we found.
What this data is and how we gathered it
Be clear about what this is. It is not a survey of fleet managers, and we are not attributing any figure here to a trade association or a research firm. It is a structured log of how three AI engines answered a fixed set of vendor-research prompts, built from the questions fleet-side contacts at our clients told us they ask, plus the questions we see in fleet forums and Reddit threads. We ran the set on a schedule across ChatGPT with search enabled, Perplexity, and Google AI Mode, and for each answer we recorded the vendors named, their order, the sources cited or linked, and whether the answer asked a clarifying question.
The prompt set covers four vendor categories: fleet maintenance providers (mobile and shop-based, including tire and PM programs), telematics and ELD platforms, fleet leasing and financing companies, and commercial truck dealers for new and used units and parts. The set includes prompts with and without geography, with and without fleet size, and with and without a stated pain or exclusion. We report shares of answers, not counts, and we round to whole numbers. Treat every figure as a shape. The shapes were stable across the period; the exact numbers moved week to week as the engines changed, and they will keep moving.
| ELEMENT | WHAT WE DID | LIMITATION |
|---|---|---|
| Engines | ChatGPT with search, Perplexity, Google AI Mode | Claude and other engines not in the set; AI Overviews logged only where they appeared for the same query |
| Prompt set | Fixed set across four vendor categories, four constraint types, with and without geography | Prompts written by us from client and forum patterns, not sampled from real user sessions |
| Logged per answer | Named vendors and order, cited sources by type, clarifying questions, geographic specificity | Source attribution in ChatGPT answers is partial; we logged what was shown |
| Reporting | Shares of answers, rounded, directional | Not weighted by real query volume; not a market-share measure |
Why fleet managers research vendors in AI engines now
Two things pushed vendor research into AI engines faster for fleets than for most B2B buyers. The first is cost pressure. The American Transportation Research Institute's 2026 update to its operational cost analysis frames the year as one where freight rates have started to recover while industry-wide costs keep accelerating, which means fleets are still under orders to hold cost discipline. A fleet manager told to cut maintenance spend by a tenth does not have three weeks to run an RFP. She wants a shortlist today, and an AI engine produces one in a minute.
The second is that fleet vendor questions are unusually well suited to an AI answer. They are constrained, operational, and comparative. A fleet manager rarely asks who the best telematics company is. She asks which ELD platforms integrate with her specific TMS, handle a mixed fleet of Class 8 tractors and Class 5 service trucks, and do not lock her into a five-year contract. That is a question a search results page answers badly and an AI engine answers directly, by pulling the constraints apart and checking each vendor against them.
The prompts we logged reflect that. Below are the shapes we saw most, by category. The specific wording varies; the structure does not.
| CATEGORY | TYPICAL PROMPT SHAPE | CONSTRAINTS THAT APPEAR MOST | FOLLOW-UP PROMPT THAT USUALLY COMES NEXT |
|---|---|---|---|
| Fleet maintenance | Mobile or shop maintenance providers for a 60-truck mixed fleet in [state] that can do DOT inspections and PMs on site | Fleet size, vehicle class mix, geography, on-site capability | Which of these have the best reviews from fleets my size |
| Telematics and ELD | ELD and telematics platforms for a 200-truck carrier that integrate with [TMS], no long contract | Integration, contract length, mixed fleet, driver-facing app quality | Compare [vendor A] and [vendor B] on pricing and support |
| Leasing and financing | Full-service lease vs finance for 25 Class 8 tractors, which companies work with a carrier at our size | Fleet size, lease type, credit or age of company, maintenance inclusion | What are the downsides of full-service leasing for a small fleet |
| Truck dealers | Freightliner or International dealers near [metro] with used Cascadias under 400k miles and a parts counter open Saturday | Brand, proximity, inventory spec, parts and service hours | Which of these dealers has the best service department reputation |
Notice the follow-up column. In most sessions the first prompt produces a shortlist and the second prompt is a trust test on the shortlist. That two-step pattern is identical to what we see when shoppers ask which car dealer to trust, and it means a vendor has to survive two different evidence checks: one for fit and one for reputation.
How the four vendor categories behave in AI answers
The first thing the logs show is that the engines are far more willing to name vendors in some categories than in others. For truck dealers and telematics platforms, the answer almost always contains specific company names. For maintenance providers and leasing companies, the engines frequently answer with generic guidance, a description of what to look for, and a suggestion to search a directory, without naming anyone. That is a vendor-side failure, not an engine-side one, and it is the most actionable finding in the data.
Share of answers that named at least three specific vendors, by category (our prompt logs, directional)
Why the gap? Truck dealers are entities with addresses, OEM affiliations, and inventory that lives on TruckPaper and Commercial Truck Trader in structured form. Telematics vendors are software companies with comparison pages, review-site listings, and integration directories that state exactly what they do. Both categories are easy to describe, so the engines describe them. Maintenance providers and leasing companies, by contrast, tend to have sites that describe a philosophy rather than a service area, a fleet-size range, a vehicle-class capability, or a contract structure. The engine cannot check them against the prompt's constraints, so it does not name them.
| CATEGORY | WHO GETS NAMED MOST | WHAT THE ENGINE CITES FOR THEM | COMMON REASON A VENDOR IS ABSENT |
|---|---|---|---|
| Truck dealers | Franchised dealers of the brand in the prompt, nearest the metro, with inventory pages the engine can read | Dealer site inventory and location pages, TruckPaper and Commercial Truck Trader listings, Google reviews | Group-level site with no dealership location pages; inventory behind a search box with no crawlable listing pages |
| Telematics and ELD | Platforms with integration directories and third-party reviews | Review sites, integration pages, trade-press comparisons, Reddit threads from drivers and dispatchers | No stated integration list; pricing and contract terms absent from the site |
| Leasing and financing | National lessors with published program pages; regional players rarely | Vendor program pages, trade press, association directories | Site describes benefits, not fleet-size ranges, lease structures, or geography |
| Fleet maintenance | National mobile networks and a few regional shops with service-area pages | Vendor service-area pages, Google reviews, fleet forum mentions | No service-area page, no vehicle-class statement, no fleet-size statement, no reviews from fleets |
The dealer column is the one closest to work we already document. The playbook for getting a truck dealer's dealership locations, inventory, and parts counters into these answers is the commercial truck dealer SEO playbook, and most of it is about making the dealer as legible as its TruckPaper listings already are.
What the engines cite when they shortlist fleet vendors
Across all four categories, the sources the engines cited fell into a narrow set of types. We grouped them and computed each type's share of all citations logged. Two things stand out: how much weight third-party review and directory sites carry, and how small the share for vendors' own sites is in the two categories where vendors are most often absent.
Share of citations by source type across all four categories (our prompt logs, directional)
The vendor's-own-site share splits sharply by category. For truck dealers and telematics it is the largest single source type, because those vendors publish pages with checkable specifics. For maintenance and leasing it is a minority, because the engines cannot find the specifics on the vendor's site and go to trade press or directories instead. The way to raise your share of citations is not to earn more links to the homepage. It is to publish the pages the engine is looking for and cannot find.
Reddit and forum citations are worth a separate note because they punch above their weight on the trust-test follow-up. When a fleet manager asks which of these has the best reputation, the engines lean on driver and dispatcher threads more than on any other source. Those threads are also where the harshest, most specific vendor complaints live. A vendor with no presence in them is judged by whatever the loudest thread says.
The anatomy of a fleet vendor prompt
The engines' behavior is easiest to understand by taking a prompt apart. Here is a composite of the maintenance-category prompts we logged, followed by the checks the engines appear to run before naming anyone.
“We run 85 trucks, mostly Class 8 day cabs plus about 20 Class 5 and 6 service trucks, out of two yards in central Ohio. Looking for a maintenance provider that can do PMs and DOT inspections on site at both yards, handle the mixed fleet, and give us actual uptime reporting. Not interested in [national network] again. Who should we talk to?”
The same anatomy applies with different nouns in the other categories. Telematics prompts turn on integrations and contract terms. Leasing prompts turn on fleet size, lease structure, and whether maintenance is included. Dealer prompts turn on brand, proximity, inventory spec, and parts and service hours. In every case the engine is running the prompt's constraints against text it can find, and the vendors that publish that text are the vendors that get named. This is the same mechanism we described for automotive at large in our generative engine optimization overview, applied to a buyer with more constraints and less patience.
What separates vendors that get named from vendors that do not
We compared vendors that were named consistently across prompts in their category against vendors we know are competitive on the ground but rarely or never appeared. The differences were not about company size or domain authority. They were about a short list of page types and evidence types.
None of those six is expensive. All six are the kind of work vendor marketing teams skip because it does not look like marketing. It looks like filling out a profile. That is the point: an AI engine is, functionally, a very fast analyst filling out a comparison grid, and it can only fill in the cells it can find.
What this means for vendors selling to fleets
If you sell maintenance, telematics, leasing, or trucks to fleets, the practical program has four parts, and the order matters.
First, publish the constraint pages. Service area, fleet size and class range, capabilities in nouns, integration lists, lease structures, inventory that is crawlable at the unit level. Each of those is a page or a section, written in plain sentences, that answers the fit half of a prompt. For truck dealers this overlaps almost entirely with the inventory and dealership location work that also wins organic search, which is why we treat commercial truck dealerships as a single program rather than SEO plus GEO.
Second, build the fleet-specific reputation corpus. Ask fleet customers for reviews and ask them to say what they run and what you did. Respond with specifics. Get listed on the category review sites and association directories, and make sure every listing agrees with the site. Third, show up in the threads under a real name, answering technical questions without pitching. Fourth, measure: a fixed prompt set for your category and geography, run weekly, logging whether you are named and what gets cited. Our AI citation tracking does this for every client, and the first month of data usually tells a vendor exactly which of the six page types it is missing.
Carriers are on both sides of this. A fleet's own managers research vendors this way, and the same fleet is being researched by drivers deciding where to apply and by shippers deciding who to tender to. The driver side has its own dynamics, covered in our piece on driver recruiting search for carriers, but the mechanism is identical: the engine names the carrier it can describe. Fleets that want to be found by drivers and shippers should read the six-item list above as applying to themselves, and the industry program for commercial trucking fleets starts there.
One caution about what not to do. Several vendors we tracked responded to being absent from AI answers by publishing long articles about the future of AI in fleet management. Those articles were never cited for a vendor prompt, because vendor prompts are not asking about the future of anything. They are asking who can do PMs at two yards in central Ohio. Write that page.
Cite this research
If you want the prompt set for your own category and geography, or want to see how your company scores against the six page types before you build anything, that is the first step of our generative engine optimization engagements for fleet vendors, and it takes about two weeks to produce a baseline. The gap between being described and being named is usually one quarter of unglamorous page work. Most of your competitors have not started it.
See where you are cited today
A free snapshot audit of your rankings and AI citations before we ever talk.
Tyler leads SEO and generative engine work for dealerships, truck dealers, fleets, and automotive software companies at Something Inc., getting them ranked and cited across every major AI engine.