• AI visibility

Guide · AI search

How do analytics and BI software companies get shortlisted when buyers ask AI?

By being named, and described accurately, when data leaders ask AI assistants to compare platforms, suggest alternatives or solve a specific reporting problem, then winning the trial on the buyer’s own data. Analytics buyers now use AI every day and shortlist very few products, so the first list matters. It also matters that AI is changing the product itself: buyers increasingly ask whether a BI tool can answer questions in plain language accurately, and vendors need public proof that theirs can.

By Underneath Team Guide Updated

The short version

  1. Data teams already work through AI: 80% of analytics professionals used AI in their daily workflow in dbt Labs’ 2025 survey (opens in a new tab) of 459 practitioners, up from 30% a year earlier, mostly through ChatGPT, Claude and Gemini.
  2. Plain-language analytics is the new buying question: 30% of those teams use AI to answer data questions in natural language, and another 29% want to but don’t yet.
  3. Incumbents bundle hard: Microsoft reported more than 40,000 paid Fabric customers (opens in a new tab), up more than 60% in a year, and says Power BI (opens in a new tab) has 30 million monthly active users.
  4. Data readiness decides deals: 63% of organizations lack, or are unsure they have, the right data management practices for AI, according to a Gartner survey (opens in a new tab) of 1,203 data management leaders.
  5. Winning accounts expand: the data platform Snowflake (opens in a new tab) reported 828 customers spending more than $1 million a year and a net revenue retention rate of 126% in mid-2026.

Who signs for a BI platform, and how much is an account worth?

A data leader and a business sponsor buy it together, usually after a trial; good accounts grow.

The buyer is split in two. A chief data officer, head of analytics or data engineering lead judges the technology: connectors, the semantic layer, governance, performance and cost. A business sponsor, often in finance, sales operations or marketing, judges whether people will use it. Budgets are rising after a lean period: in the dbt survey, 30% of respondents reported budget increases for their data teams, against just 9% the year before, and 40% reported headcount increases. EdTech shows a similar split between the people who use a tool and the people who approve it, with teachers finding tools that districts then buy, as our guide to EdTech buyers who ask AI explains.

Two market facts shape every deal:

  • Bundling. Microsoft has been named a Leader in Gartner’s Magic Quadrant for analytics and BI platforms for the eighteenth consecutive year, and Power BI is now part of Fabric. Many buyers already own a BI tool through a bigger contract, so an independent vendor must win a comparison against something that looks free.
  • AI is moving the category. In a 2024 survey of 1,000 data and business leaders by MIT SMR Connections, sponsored by ThoughtSpot (opens in a new tab), 67% were already using generative AI for an analytics use case. Vendor-sponsored, but consistent with the dbt data.

What a customer is worth depends on how it pays. Seat-based BI grows with users; consumption-based platforms grow with data and queries. Snowflake’s 126% net revenue retention means, roughly, that its existing customers spent 26% more than a year before, and it added 692 net new customers in the quarter. Snowflake is a data platform chosen by data leaders rather than a BI tool, but it shows how analytics accounts compound once they become part of daily work.

Where does AI already sit in the analytics buying journey?

Inside the data team’s daily work, and increasingly at the start of software research.

Analytics buyers are heavy AI users. dbt found 70% of respondents use AI for analytics development in some form, mainly through general-purpose assistants such as ChatGPT, Claude and Gemini. A data engineer who asks an assistant to write a query today can just as easily ask it which tool to use for the next project. That makes the assistant both a research channel and, for simple questions, a competitor to the product, we infer.

Surveys of technology buyers beyond the data team point the same way. In TrustRadius’s 2026 report (opens in a new tab) on nearly 2,500 technology buyers and vendors, 63% of buyers used AI during their purchase journey and 94% of them fact-checked its responses at least some of the time. These figures cover technology purchases in general, not BI alone.

Google’s AI features sit on top of BI research as well. Of the eight industries in our study of 1,248 US searches, B2B software and technology had the highest rate: an AI Overview appeared on 96.0% of its searches.

Which questions do analytics buyers ask AI assistants?

Comparisons, alternatives, use cases, accuracy and cost. We wrote the data-team prompts below to illustrate the pattern; they are not records of real buyer questions.

StageIllustrative prompt
Use case“What is the best way to give our sales team self-serve dashboards on top of Snowflake?”
Alternatives“What are the main alternatives to Tableau for a company standardizing on Google Cloud?”
Comparison“Power BI or Looker for a 2,000-person company with a dbt semantic layer?”
Natural language“Which BI tools can answer plain-English questions accurately without exposing raw data?”
Embedded“What embedded analytics platforms work for a multi-tenant SaaS product?”
Cost“How does consumption pricing compare with per-seat BI licensing at 500 users?”
Governance“Which BI platforms support row-level security and a central metrics layer?”

Behind each of these prompts the assistant may run searches of its own. Google says AI Overviews and AI Mode may use a “query fan-out” technique (opens in a new tab), which issues multiple related searches, and OpenAI’s help center says ChatGPT search typically rewrites a question (opens in a new tab) into targeted queries. Those searches often go looking for outside judgment: our hidden-searches study found that 43.8% of ChatGPT’s answers involved a search aimed at a named publication, ranking or award. For BI, a reasonable expectation is that analyst evaluations, review sites and practitioner blogs are among those named sources.

How does a mention in an AI answer become a BI trial, then a contract?

Through a short shortlist and a trial on the buyer’s data, then expansion as usage spreads.

  1. A data leader or analyst asks an assistant a use-case, alternatives or comparison question.
  2. The answer names a few platforms and cites sources: reviews, comparisons, documentation, analyst coverage.
  3. The buyer narrows to a short list. Those lists are short: 83% of buyers shortlisted three or fewer products, TrustRadius reports. In the Gartner Digital Markets survey of 3,500 software buyers, initial lists averaged 4.4 options, and 81% of buyers end up buying from that initial list most or all of the time.
  4. A trial or proof of concept runs on real data. In the same Gartner Digital Markets survey, 62% of buyers said the trial is their top factor in the final decision.
  5. A contract follows, priced by seats or consumption, and grows as more teams use it.

AI answers matter most at steps 1 to 3. For analytics vendors, step 4 is where a vendor proves the claims the answer repeated: accuracy, speed, governance. A vendor that is described wrongly, for example with outdated pricing or a missing connector, may never get to run that trial.

Why does an assistant name one BI platform and not another?

No platform documents how it chooses; studies point to third-party sources, and data leaders reward proof they can check.

What Google and OpenAI say. Both describe AI answers that search the web and cite sources; neither says how a BI platform ends up on the list.

What research has measured. On US software questions, earned sites supplied 72.7% of AI search sources, against 45.4% for Google, according to Chen and colleagues (opens in a new tab). In our four-assistant study, assistants agreed most on B2B software, with an overlap of 0.543 on a scale from 0 to 1, still far from full agreement. Lists also drift between runs of the same question: our consistency study found only 25.2% of the brands ChatGPT named held their place across all five.

What analytics buyers trust. Several trust factors are specific to the category:

  • Accuracy of AI answers on data. Of the dbt respondents using AI for natural-language questions, two-thirds rely on plain query generation and one-third on a semantic layer, which past research links to higher accuracy. 27% plan to increase investment in semantic layer tooling. Vendors with published accuracy methods give buyers something to test.
  • Data quality and readiness. Poor data quality is the challenge data teams cite most, at over 56% in the dbt survey, and Gartner predicts organizations will abandon 60% of AI projects unsupported by AI-ready data through 2026. The same governance check shapes how data science platforms win enterprise buyers.
  • Reviews and peers over analyst reports. TrustRadius found 74% of buyers use reviews, while analyst reports were used by only 13%, a 63% decrease since 2022. Analyst recognition still appears in vendor marketing, but buyers lean more on peers.
  • Governance and security. Row-level security, auditability and where data is processed are standard checks for any platform that touches company data.

Our inference. The proof analytics buyers check is mostly public: documentation, connector lists, pricing pages, reviews, benchmarks and practitioner write-ups. We would expect a BI vendor whose connector lists, pricing examples and accuracy methods are public, specific and consistent to give AI answers more to cite, but nobody has tested that for analytics software.

What does a BI vendor lose when assistants leave it out?

A place on a list of three or four, which is most of the deal.

  • Short lists leave little room. With 81% of software buyers purchasing from their initial list most or all of the time, a vendor absent from AI-shaped lists has to displace a product the buyer already favors.
  • The bundled default wins by silence. If an answer names only the tool the buyer already owns, an independent vendor loses without a comparison, we infer.
  • Wrong facts end trials early. Outdated pricing models, connectors or AI features in an answer can remove a vendor from consideration. Correcting a stale connector list or pricing model is covered in how to fix wrong brand information in AI answers.
  • Regret is a renewal risk and an opportunity. Gartner Digital Markets found 59% of buyers regret at least one software purchase from the past 18 months. Buyers looking to replace a tool will ask for alternatives, and the answer decides who gets the call.

How does GEO work for an analytics or BI company?

It makes your proof on accuracy, governance, connectors and cost easy for AI assistants to find and repeat. It cannot promise any BI vendor a spot on a data leader’s shortlist.

  1. One clear identity. Say what the platform is for (self-serve BI, embedded analytics, natural-language analytics, a metrics layer) and for whom, the same way on your site, review profiles and partner marketplaces.
  2. Public technical proof. Publish connector lists, security and governance documentation, pricing models with examples, and how your natural-language features stay accurate. Write for the data engineer who will check it.
  3. Honest comparison and alternatives pages. Buyers ask “X or Y” and “alternatives to Z.” Answer with real trade-offs, including where the bundled option is good enough; see whether comparison pages help.
  4. Reviews and practitioner coverage. Encourage detailed reviews from data teams and earn coverage in data engineering publications and communities. Third-party lists tend to outweigh your own pages, as we show in which pages to target.
  5. Authority beyond your site. Talks, partner listings with data platforms, and independent benchmarks build the record assistants draw on; see how brands build authority for AI search.
  6. Measure across assistants and runs. Repeat your BI questions in ChatGPT, Google AI Overviews and AI Mode, Gemini, Perplexity, Copilot and Claude; how many prompts to track explains how many to use. The broader approach for subscription software sits in B2B SaaS revenue from AI search.

What is still unknown about AI search in BI buying?

Nobody has measured how often data leaders choose BI tools through AI, or whether it lifts revenue.

  • No BI-specific buyer study. The TrustRadius and Gartner Digital Markets figures cover software buyers in general.
  • Many sources are vendors. dbt Labs, Microsoft, Snowflake and ThoughtSpot all sell in this market; TrustRadius and Gartner Digital Markets sell vendor visibility.
  • AI may change what buyers buy. If assistants answer simple data questions directly, demand may shift toward governed data and semantic layers rather than dashboards. That is our inference, not a finding.
  • Revenue effects are untested. Whether a mention leads to more analytics trials or contracts is open; see does AI visibility drive business results.

What should a BI vendor check before buyers start their next trials?

Run your buyers’ comparison and alternatives questions through the assistants, then close the proof gaps you find.

Start from three people: the data leader, the analyst and the business sponsor. Ask their questions of each main assistant more than once, and log who is named, which sources are cited, and whether your pricing model, connectors, governance and natural-language features come out right. A miss usually traces back to proof that is not public, or to thin coverage from reviewers and practitioners.

That review is where we start with analytics vendors: talk to us about an audit of your BI shortlist visibility. We will show which comparison and alternatives answers include or omit you, and which public proof gaps are most likely costing you trials, proofs of concept and seat or consumption growth. Our generative engine optimization service page explains how that audit becomes a program of connector, governance and pricing proof, with comparison pages and practitioner coverage.

Frequently asked questions

Do data teams use ChatGPT to choose BI tools?

They use it heavily for work: 70% of dbt’s respondents use AI for analytics development. No study isolates vendor selection.

Do Gartner Magic Quadrant placements still matter?

Less than before for buyers: only 13% of technology buyers used analyst reports in their decision, TrustRadius found.

How do we compete with a BI tool bundled into a larger contract?

Be specific about where you are better, and say it on pages assistants can cite: comparisons, benchmarks and reviews.

Should we publish our natural-language accuracy methods?

Yes, if you can back them. Only 30% of dbt’s respondents use AI for natural-language data questions; accuracy is the barrier.

When would AI visibility show up in BI trials and contracts?

Expect trials first and contracts later; enterprise analytics deals follow a proof of concept on real data.

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