Spend management software in the US: which platforms ChatGPT, Gemini and Google AI Mode recommend

We asked ChatGPT, Gemini and Google AI Mode 12 questions finance and procurement leaders ask about spend management: source-to-pay, invoice matching, supplier risk, treasury. The answers name 30 platforms. Coupa and SAP Ariba lead almost everywhere, and a newer name, Ramp, wins a clearly defined part of the field.

Coupa and SAP Ariba set the list

Coupa is named in 33 of the 36 answers, 11 from each model, and first in 18. SAP Ariba follows with 27. Ivalua is next with 16, then GEP with 12 and Jaggaer and Oracle's procurement cloud with 10 each. The questions were written from Coupa's offer (see the method), which helps it; the order of the others is the more telling result.

On the 4 questions about enterprise source-to-pay, both Coupa and SAP Ariba are in all 12 answers and Ivalua in 9. The category looks settled when the buyer is described as a large company.

1. Coupa33 · 92%
2. SAP Ariba27 · 75%
3. Ivalua16 · 44%
4. GEP12 · 33%
5. Oracle Procurement10 · 28%
6. Jaggaer10 · 28%
7. Workday8 · 22%
8. Zycus7 · 19%
9. Basware3 · 8%
10. Icertis3 · 8%
11. Kinaxis3 · 8%
12. Blue Yonder3 · 8%
13. Sievo2 · 6%
14. o92 · 6%
15. EcoVadis2 · 6%

Ramp owns spend and cash, and nothing else

On the 4 questions about spend visibility, expenses, treasury and cash flow, Ramp is named in all 12 answers, level with or ahead of Coupa. It is first in 4 answers overall. It is named in none of the other 24 answers, on enterprise procurement, sourcing and supply chain.

Brex follows the same pattern, with 6 answers on the same ground. Treasury brings specialists: Kyriba is named in 4 answers of 6 and Centime in 3. SAP Concur is named in 6 answers, mostly about expenses.

1. Ramp12 · 33%
2. Brex6 · 17%
3. Procurify6 · 17%
4. Zip6 · 17%
5. SAP Concur6 · 17%
6. Kyriba4 · 11%
7. Tipalti3 · 8%
8. Centime3 · 8%
9. Payhawk2 · 6%
10. Precoro2 · 6%

Also named, in fewer than 2 answers each: 5 other spend, expense and treasury tools.

ChatGPT, Gemini and Google AI Mode, October 26, 2026. One answer can name several spend, expense and treasury tools.

Supply chain and risk bring other names

On supply chain design and planning, Kinaxis, Blue Yonder, Coupa and Oracle are named by all three models. On supplier risk and ESG, the lists are the shortest of the study: Gemini names only EcoVadis, ChatGPT SAP Ariba and EcoVadis, Google AI Mode SAP Ariba and Coupa.

The models favour different suites

Gemini names 5.9 platforms per answer, ChatGPT 5.6 and Google AI Mode 4.8. Eighteen of the 30 are named by all three. Oracle's procurement cloud is named in 7 of ChatGPT's answers, against 2 for Gemini and 1 for Google AI Mode. Workday goes the other way, with 5 answers from Gemini and 1 from ChatGPT. Basware is named only by ChatGPT.

Named in 3 answers or more	ChatGPT	Gemini	Google AI Mode
Coupa	11	11	11
SAP Ariba	9	10	8
Ivalua	7	4	5
Ramp	4	4	4
GEP	3	4	5
Oracle Procurement	7	2	1
Jaggaer	3	4	3
Workday	1	5	2
Zycus	2	2	3
SAP Concur	2	2	2
Brex	1	3	2
Procurify	2	3	1
Zip	2	3	1
Kyriba	1	1	2
Kinaxis	1	1	1
Tipalti	1	1	1
Centime	1	1	1
Blue Yonder	1	1	1
Basware	3	·	·
Icertis	2	·	1

Analysts carry weight here

coupa.com is the website cited most, in 13 answers, and gartner.com is second, in 12. sap.com follows with 8, then ciopages.com and ivalua.com with 7 each, and ramp.com with 6. ChatGPT cites 8.5 sites per answer, Gemini 3.2 and Google AI Mode 2.3.

Enterprise suite websiteFinance tool websiteOther website

1. coupa.com13 · 36%
2. gartner.com12 · 33%
3. sap.com8 · 22%
4. ciopages.com7 · 19%
5. ivalua.com7 · 19%
6. ramp.com6 · 17%
7. oracle.com5 · 14%
8. emburse.com4 · 11%
9. jaggaer.com4 · 11%
10. brex.com3 · 8%

What this means for spend and procurement software

• Analyst coverage feeds the answers. Gartner is cited almost as often as the leader's own site. For an enterprise vendor, analyst reports are part of what assistants read.
• A narrow position can win. Ramp is absent from most of the field and in every answer on its own ground. A challenger can choose its questions rather than fight on all of them.
• Say who you are for. Mid-market and enterprise produce different shortlists. Company size and use case need to be explicit in the pages.
• Supplier risk is thin. Few names and short answers leave room for a vendor with clear, specific content on risk and ESG.

Method

12 questions in English, written from the offer of one vendor, Coupa, across enterprise source-to-pay, spend and expenses, sourcing and suppliers, treasury and supply chain. Each question was asked once to ChatGPT (with web search), Gemini and Google AI Mode, for the US market, on October 26, 2026: 36 answers. We looked in each answer for 37 platforms (names and common spellings), counted the answers naming each, their position among those names in the answer, and the websites cited. Oracle's procurement cloud is counted under the names answers use for it, such as Oracle Fusion.

This is a snapshot: three models, one day and 12 questions. AI answers vary from one session to the next, so a difference of one or two answers between two names is not significant. Because the questions start from one vendor's strengths, they favour it and platforms like it; they do not measure the whole market.