Three products, Expensify, Ramp and Zoho Expense, were each named in all 6 of the 6 answers when we asked ChatGPT and Gemini to recommend expense management software on 3 August 2026. Brex came next, named in 5 of the 6 answers. Nothing else reached even half: the names in fifth place, BILL Spend & Expense, Certify and SAP Concur, each appeared in just 2 of 6.
We ran six buyer questions, the sort a finance manager or a marketing lead actually types: the plain "best expense management software," then "for small teams," "what should I use," "best free," "most affordable," and a catch-all "recommendations." For each one we logged every product ChatGPT and Gemini named and every web page the answers leaned on. Fifteen product names came up across the 6 answers, but only four of them appeared in more than a third of those answers. The AI shortlist for this category is really a top four with a long tail of names that surfaced once or twice.
How the shortlist changed from question to question
The four constants did most of the work, but each phrasing of the question pulled in a different supporting cast, and this is the part worth watching. The bare "best expense management software" query returned the widest list, 8 products, and four of them, Coupa, Concur Expense, Navan and Rippling, appeared in that one question and nowhere else across the 6.
The "best free expense management software" question rewrote the list. It brought in Everlance and TravelBank, two names that showed up in that single question only, and it was the first place BILL Spend & Expense and SAP Concur appeared at all. It was also the one answer of the 6 that dropped Brex completely, which fits a corporate-card product that has no free tier.
"Most affordable expense management software" did the opposite thing. It collapsed to just 4 names, and every one of them, Zoho Expense, Expensify, Ramp and Brex, was already in the top four. The affordability question surfaced no new product at all, it just reprinted the leaders.
Two more names lived in exactly one question each: Mercury appeared only under "best for small teams," and OFX appeared only under "recommendations." Add it up and 8 of the 15 names, Coupa, Concur Expense, Navan, Rippling, Mercury, Everlance, TravelBank and OFX, depended on the wording of a single question. Change the question and more than half the list changes with it. The four that hold steady across all six phrasings are the only names a vendor can count on.
The pages behind the answers
The most-cited source by a wide margin was Reddit, which fed 6 of the 6 answers. A community discussion board, not a vendor and not a directory, was the one source present in every single answer. Quora added more community threads to 3 of 6.
Two research and review directories came next: G2 was cited in 5 of 6 answers and Gartner in 5 of 6, with the roundup publisher PCMag in 3 of 6. The familiar "best expense software" listicle and review-grid pages are clearly doing real work. One big directory was missing: Capterra was cited in 0 of the 6 answers, despite being one of the largest review sites in this space. In this run, G2 counted and Capterra did not.
Here is the finding that matters most if you sell in this category: vendor-owned domains fed the answers, and the vendors whose own pages were cited were the vendors who got named. Ramp.com was cited in 3 of 6 answers and Ramp was named in all 6. Expensify.com was cited in 2 of 6 and Expensify was named in all 6. Brex.com was cited in 4 of 6 and Brex was named in 5. Zoho.com and OFX.com each fed 1 answer, and both Zoho Expense (6 of 6) and OFX (1 of 6) were named. The assistants were not only reading about these companies on third-party sites, they were reading the companies' own pages and repeating them back.
One vendor domain broke the pattern, and it is the most useful line in the data. Mercury.com was cited in 4 of 6 answers, more often than Ramp's or Expensify's own sites, yet Mercury the product was named in only 1 of 6. Mercury's pages were being used as a source about the category, not as a recommendation for Mercury. Getting cited and getting recommended are two different outcomes, and this is the clearest example of the gap.
What this means if you sell expense management software
If your product is in this category and you want ChatGPT and Gemini to name it, the data points at a concrete list, and none of it is "get more market share." The names that won were the names present in these exact places: Reddit threads (6 of 6), G2 and Gartner (5 of 6 each), PCMag roundups (3 of 6), and their own clear vendor pages, since Ramp, Expensify, Brex and Zoho all fed citations from domains they own.
Size is not what decided it. Certify and BILL Spend & Expense are established products and each landed in only 2 of 6 answers, while Zoho Expense reached 6 of 6. The assistant builds its shortlist from whichever pages it can pull and trust on the day of the search, so the real question is not "are we a market leader," it is "when a buyer runs this search, do our pages and our mentions come back in the answer." In practice the work is getting into the pages an assistant retrieves for this search, the Reddit threads, the G2 and Gartner grids, the roundups, and your own site, rather than buying your way up a market-share chart. That is measurable, and it moves.
The full list, counted
| Product | Named in | Share |
|---|---|---|
| Expensify | 6 of 6 | 100% |
| Ramp | 6 of 6 | 100% |
| Zoho Expense | 6 of 6 | 100% |
| Brex | 5 of 6 | 83% |
| BILL Spend & Expense | 2 of 6 | 33% |
| Certify | 2 of 6 | 33% |
| SAP Concur | 2 of 6 | 33% |
| Concur Expense | 1 of 6 | 17% |
| Coupa | 1 of 6 | 17% |
Two of those rows are the same company. SAP Concur was named in 2 of 6 answers under that label and once more as "Concur Expense" in 1 of 6, and we left them as separate rows because that is how the assistants wrote them, rather than merge them into a combined number the measurement never produced. It is a small reminder that these tools do not reach for tidy canonical product names, they repeat whatever string came back from the page.
Where Glotier fits
Glotier does not sell expense management software, so we are correctly absent from all 6 of these answers, and we would be concerned if we were not. Measuring this is the thing we actually do. This whole page is one run of the check we point at a customer's own category: the same six-question fan-out, the same log of who got named and which pages fed each answer.
If you sell in a category like this one, you can see your own version of this table. The check is free, there is no account, and it takes about a minute.