• AI visibility

Guide · AI search

Are AI assistants now shaping which enterprise software gets shortlisted?

Increasingly, yes: enterprise buyers now use AI assistants to gather information on vendors before they talk to any of them, and many do it inside the company’s own Copilot or ChatGPT workspace. For a vendor selling six- and seven-figure contracts, the prize is not website traffic but a place on the early list that a buying committee then tests for months. That place depends mostly on what independent sources say about you, and buyers still check it with people before they sign.

By Underneath Team Guide Updated

The short version

  1. Enterprise buyers lean on AI more than smaller firms: in G2’s 2025 survey (opens in a new tab), buyers at companies with 1,000 to 5,000 employees named review sites (56%) and AI search (55%) as their top two research sources, against 38% and 35% for small businesses.
  2. Much of that research happens behind the firewall: Forrester (opens in a new tab) found 61% of business buyers use private AI tools provided by their organization, and they are four times as likely as consumers to use Microsoft Copilot.
  3. AI is a starting point, not the final word: in Gartner’s (opens in a new tab) survey of 645 B2B buyers, 45% used generative AI in a recent purchase, mainly to gather information on vendors, and 69% prefer to validate AI-generated insights with sales reps.
  4. The deals are large and slow: Forrester (opens in a new tab) found that 78% of buyers making purchases of $10 million or more run a trial first, and in North America the average buying cycle in 6sense’s 2025 study (opens in a new tab) was 11.1 months.
  5. One enterprise customer can be worth millions a year: ServiceNow ended the second quarter of 2026 with 658 customers paying more than $5 million in annual contract value.

Who signs an enterprise software contract, and what can one account pay?

A large committee buys it, with procurement involved from the start, and one won account can pay millions a year.

Enterprise software is the slice of business software sold to large organizations: platforms for IT, finance, HR, data and customer service, sold through sales teams rather than a credit card form. The overall market is growing fast. Gartner (opens in a new tab) forecast worldwide software spending of $1,433,037 million in 2026, up 15.2%, and said generative AI features “are now ubiquitous across software already owned and operated by enterprises.”

The buyer is a group, not a person. Forrester’s 2026 study (opens in a new tab) found that 13 internal stakeholders and nine external participants influence the average business purchase, and that procurement staff are decision-makers in 53% of buying cycles. When the purchase includes generative AI features, the buying group doubles, from seven members to 14.

Public filings show what one enterprise customer is worth:

CompanyDisclosed figure
ServiceNow, Q2 2026658 customers with more than $5 million in annual contract value; 123 deals over $1 million in net new annual contract value in the quarter
Snowflake (opens in a new tab), quarter to July 31, 2026828 customers with more than $1 million in trailing 12-month product revenue; net revenue retention of 126%

Net revenue retention of 126% means existing customers, as a group, spent 26% more than a year earlier. In enterprise software the first contract is often the smallest one, so winning a seat at the first evaluation can be worth far more than the opening deal.

This article is about those large, committee-led deals. For self-serve and mid-market software, where trials and seat-based plans drive revenue, see how B2B SaaS companies generate revenue from AI search. Health technology sold to employers and health plans is covered in our guide for healthtech vendors.

Where do AI assistants sit in an enterprise software evaluation?

At the research and shortlist stage, often inside tools the buyer’s employer provides, and before sellers are contacted.

The survey evidence is consistent across firms:

  • In Forrester’s Buyers’ Journey Survey of nearly 18,000 global business buyers, 94% reported using AI during their buying process.
  • Much of it happens in tools the employer supplies: according to Forrester, 61% of business buyers, a majority, use private AI tools their organization provides. They are twice as likely as consumers to use ChatGPT and four times as likely to use Microsoft Copilot, and more than half use private versions behind their firewall.
  • In G2’s 2025 survey of 1,100 decision-makers, AI chatbots were the single source that most influenced vendor shortlists, at 17.1%, ahead of review sites (15.1%), vendor websites (12.8%) and market research firms (10.6%).
  • In Gartner’s survey, buyers used an average of seven information sources during a recent purchase.

The private-tool finding matters for how visibility works. Microsoft documents that when web search is used, Microsoft 365 Copilot (opens in a new tab) “generates a search query that it sends to the Bing Search service.” OpenAI documents that ChatGPT search (opens in a new tab) “typically rewrites your query into one or more targeted queries” that it sends to search providers. So an enterprise buyer’s private assistant still reads the public web; our inference is that a vendor’s visibility in Bing deserves as much attention as in Google.

Analyst research is changing too. In TrustRadius’s 2026 survey (opens in a new tab) of technology buyers, analyst reports were used by only 13% of buyers to make their purchase decisions, a 63% decrease since 2022. Gartner itself says clients now use its own AI tool, AskGartner. We infer that analyst opinion now reaches many buyers through AI summaries rather than full reports.

Which questions do enterprise buyers put to AI assistants?

Questions about fit at scale, security, integrations, total cost and proof from peers. We wrote the committee-style prompts below as illustrations; they were not captured from real enterprise buyers.

ConcernIllustrative prompt
Category at scale“Which IT service management platforms are used by global banks with more than 50,000 employees?”
Alternatives“What are the alternatives to Workday for a manufacturer running SAP?”
Comparison“Snowflake vs Databricks for a regulated insurer: governance, cost and lock-in?”
Security and compliance“Which customer data platforms are FedRAMP authorized and support EU data residency?”
Integrations“Does this procurement suite integrate with Oracle Fusion and Coupa?”
Scalability and risk“Has this vendor had major outages, and what do its uptime commitments say?”
Proof“Which Fortune 500 companies use this platform, and what do reviewers say about implementation?”
Cost“What does a 5,000-seat deployment of this platform typically cost over three years?”

Each of these questions sends the assistant to look for proof. When we logged the searches behind ChatGPT’s answers to 80 buyer questions in our hidden-searches study, 43.8% of answers had gone looking for a named publication, ranking or award. In enterprise software, we infer, those named sources include analyst evaluations, peer-review platforms and industry press.

How does AI visibility turn into enterprise pipeline?

Through the first shortlist: named early, tested by a committee, proven in a trial, then won as a multi-year contract.

  1. Early research. A team member asks an assistant to map the category and summarize options. G2 found AI chatbots were the top influence on shortlists.
  2. Shortlist. Buyers narrow the field quickly. TrustRadius found 83% of buyers shortlisted three or fewer products.
  3. Validation. The committee checks the assistant’s view with people. Gartner found 69% prefer to validate AI-generated insights with sales reps, and 51% say they are more likely to encounter misleading information from generative AI, while 49% say the same of a sales rep.
  4. Proof. More than 60% of business buyers use some form of trial, and 78% of buyers making $10 million-plus purchases run one first, according to Forrester.
  5. Contract and expansion. The winner signs a multi-year deal and, as Snowflake’s retention figure shows, often grows within the account.

The value arrives at steps four and five, months after the AI answer. That is why we suggest enterprise vendors judge AI visibility by its effect on pipeline: opportunities that name an AI assistant as a source, and win rates where the vendor was on the first list. Traffic counts miss most of it, since the research happens in private tools that send no clicks.

What earns an enterprise vendor a mention in an assistant’s answer?

Mostly independent evidence the assistant can find and check; the platforms do not publish their selection rules.

Documented by the platforms: Google says AI Overviews and AI Mode may run several related searches (opens in a new tab) across subtopics, and Microsoft and OpenAI document that Copilot and ChatGPT turn prompts into web searches. None of them says how vendors are chosen for an answer.

Observed in studies:

  • Independent coverage counts most. In our brand-entity study, each tenfold increase in the number of independent sites naming a brand in the pages an assistant cited went with 4.7 times the odds of being recommended. It was the strongest predictor we measured.
  • Google rankings are only part of it. In our rankings study, only 8.3% of the pages ChatGPT cited ranked in Google’s top 10 for the question.
  • Readable sites help, according to one vendor study. In a study by ora research (opens in a new tab) across 1,056 businesses, agent-ready businesses had answers built from their own pages 78% of the time against 56%. The authors sell tools in this area.

Our inference for enterprise software: the trust factors are analyst evaluations, peer reviews from companies of similar size, named customer references, public security and compliance documents, integration listings in partner marketplaces, and clear pricing logic. For security vendors specifically, see cybersecurity software and AI search.

What does it cost to be missing from the early shortlist?

Possibly a whole buying cycle, because enterprise buyers decide their shortlist early and renew slowly.

The cost is inferred from buying behavior, not measured directly:

  • Lists form early. In 6sense’s study, 94% of buyers ranked their shortlist vendors before contacting any of them, and first contact came when buyers were 61% of the way through their buying process. A vendor absent from the AI-assisted research may never hear the deal existed.
  • Cycles are long. With North American buying cycles averaging 11.1 months, a missed evaluation can mean waiting until the winner’s contract comes up for renewal, we infer. Security teams face this at SIEM migrations, covered in our guide for SIEM vendors.
  • Buyers switch for AI. In G2’s survey, half of enterprise buyers at companies with 1,000 to 5,000 employees said they had switched vendors for better AI, as did 41% at companies with more than 5,000. Incumbents cannot assume the renewal is safe either.

How much pipeline any single vendor loses this way has not been published. Anyone quoting a precise figure is estimating.

What does GEO mean for a vendor selling large enterprise contracts?

Generative engine optimization (GEO) makes your company easy for a buying committee’s assistant to understand, verify and recommend.

For enterprise software, that usually means:

  1. Entity clarity. Consistent product names, categories and descriptions across your site, partner marketplaces, review profiles and reference sites, so an assistant does not confuse products or editions.
  2. Proof that can be checked. Public trust and compliance pages (SOC 2, ISO 27001, FedRAMP status, data residency), uptime history, and integration documentation that is readable without a login.
  3. Third-party coverage. Analyst evaluations where they apply, peer reviews from enterprise customers, industry press, and named case studies that other sites can cite.
  4. Answer-ready pages. Plain pages for enterprise questions: migration from a named competitor, total cost logic, implementation timelines, and honest comparisons. The evidence on competitor comparisons is summarized in do comparison pages help B2B brands get cited by AI.
  5. Coverage across assistants. Check ChatGPT, Copilot (and Bing behind it), Gemini, Google’s AI features and Perplexity, because your buyers’ employers choose the tool.
  6. Pipeline measurement. Add AI assistants to “how did you hear about us” fields, ask in discovery calls which sources shaped the list, and track AI visibility over repeated runs, since a single answer is only a sample.

No GEO program can guarantee that a committee’s assistant recommends you. What it can do is make sure that when a buyer’s assistant goes looking, it finds accurate, verifiable evidence about you.

What is still unknown about AI and enterprise shortlists?

Whether AI visibility wins enterprise contracts: buyers clearly research vendors with AI, but no study links it to signed deals.

  • No controlled link to revenue. No published study traces enterprise contracts back to AI answers. Of everything here, outcomes are the least proven; see does AI visibility drive business results.
  • Private tools are hard to observe. What a buyer’s company Copilot answers is not visible to vendors, and its admins may switch web search off.
  • Surveys mix segments. Forrester, Gartner and TrustRadius cover many kinds of business purchases; only some figures are broken out for large companies. G2 and TrustRadius run review platforms and have an interest in these findings.
  • Selection is a black box. Answers vary between runs and between assistants, and the platforms do not publish ranking rules.

What should an enterprise vendor check before its next large evaluation begins?

Put your buying committees’ questions to the assistants they use, then compare the answers with your pipeline.

A useful first review covers three things: whether you are named for your core categories and use cases, which competitors and third-party sources appear instead, and whether security, integration and pricing facts about you are correct. That shows which proof is missing before a committee goes looking for it.

For vendors whose pipeline depends on making the first list for multi-year contracts, we offer that review: ask us to map how assistants present you to buying committees. We will show how assistants describe you against competitors, which evidence gaps keep you off early shortlists, and what would strengthen your place before procurement gets involved. Our generative engine optimization service page covers the longer program that follows, such as public trust and compliance pages, enterprise peer reviews and checks across the assistants buyers’ employers provide.

Frequently asked questions

Do enterprise buyers use ChatGPT or Microsoft Copilot to research vendors?

Both, and often private versions at work. Forrester found 61% of business buyers use private AI tools from their organization and are four times as likely as consumers to use Microsoft Copilot.

Does a Gartner Magic Quadrant placement still matter for AI visibility?

Probably, though no study measures it directly. Analyst reports were used by only 13% of technology buyers in TrustRadius’s 2026 survey, but assistants often search for named rankings and publications, so analyst coverage can reach buyers through AI answers.

Should we measure AI search by website traffic?

Not for enterprise deals. Much of the research happens in private assistants that send no clicks, so measure shortlist presence, sourced pipeline and win rates instead.

How is this different from GEO for mid-market SaaS?

The buyer is a committee, the cycle runs close to a year, and proof matters more than price. Mid-market SaaS turns AI visibility into trials and signups; enterprise software turns it into a seat at a long evaluation.

Sources

Free strategy call

Some questions are easier to answer about your own business.

Bring the one that matters most. On a free 30-minute call we’ll take a first look at it and send you a short written read afterward.

Ask about my businessHow we work