• AI visibility

Guide · AI search

How does a banking technology vendor get onto a bank’s shortlist when the first research happens in AI?

By being named, accurately, when a bank or credit union maps the market before it writes a request for proposal. Core banking, payments and fraud platforms are bought rarely, through long evaluations that regulators expect banks to document, so the few early answers a buyer reads can shape a decision that lasts a decade. For a banking technology vendor, AI visibility is built on checkable proof: named institutions, conversion records, resilience facts and independent coverage.

By Underneath Team Guide Updated

The short version

  1. The buyer pool is small and shrinking: the FDIC counted 4,238 insured institutions in mid-2026, down 41 in one quarter, and the NCUA 4,214 federally insured credit unions.
  2. Core decisions are rare: an Independent Community Bankers of America executive estimates just 2% to 3% of institutions convert to another core each year, and only 19% of institutions in an American Bankers Association survey (opens in a new tab) said they were likely to switch at their next renewal.
  3. Each win is large: Jack Henry (opens in a new tab) called its 58 competitive core wins in fiscal 2026 a record, and 14 of them were institutions with more than $1 billion in assets.
  4. Selection starts years ahead: choosing and converting a core can take one to three years, and CUInsight (opens in a new tab) reports some vendors are already booking conversion slots into 2029.
  5. Budgets are rising: in Bank Director’s 2025 Technology Survey (opens in a new tab), 71% of banks increased technology budgets, with a median increase of 10%.

This guide is about how banking technology vendors are found and described by AI assistants. It is not legal or regulatory advice; claims about compliance, resilience or certifications should go through your own legal and risk review.

Who buys banking technology, and what is one customer worth?

A small executive group at a bank or credit union buys it; one core customer means years of revenue.

The buyer is an institution, not a person. In Bank Director’s survey of 141 directors and executives at US banks under $100 billion in assets, 54% said a management-level team or steering committee gives final approval for technology investments, 48% said a C-level executive other than the chief information or technology officer holds that authority, and just over a quarter said the board is directly involved in major technology decisions. A typical evaluation team, we infer, includes the CEO or COO, the technology lead, the chief risk officer and, for fraud and payments, the operations and fraud heads.

The market is concentrated in a few thousand institutions, and mergers keep shrinking it. In the second quarter of 2026, four banks opened while 36 institutions merged with other banks, according to the FDIC. The credit union count fell from 4,370 a year earlier. Every merger removes a potential customer and often forces a technology choice for the combined institution.

What a win is worth shows in the vendors’ own results. Jack Henry, which reports about 7,400 bank and credit union clients, recorded fiscal 2026 revenue of $2.54 billion. Its processing revenue rose to about $1.10 billion, and its faster-payments revenue grew 49.5% as institutions adopted real-time payment services. The economics, we infer, follow the same pattern across core, payments and fraud: the contract is won once and then earns for years through processing, transaction and add-on revenue.

That is a different sale from the finance-team software covered in our guide to fintech software buyers, where controllers can start a trial in a week. Here, the trial is a conversion.

How do banks choose a core, payments or fraud vendor today?

Slowly, through peers, consultants and formal proposals, often starting two or three years before a contract ends.

The trigger is usually the renewal date. The American Bankers Association’s 2024 Core Platforms Survey, with nearly 800 institutions responding, found satisfaction with core providers falls the longer a contract runs, reaching only 41% within two years of renewal. Aside from cost, the main reasons institutions gave for considering a change were disappointment with customer service (42%) and the relationship with their core provider (38%).

From there, the process is long and staged:

  1. Early market scan. CUInsight advises institutions to “survey the market and identify high-level vendor fit before committing to a full RFP process.” It recommends starting evaluations “at least three years before contract expiration” because popular vendors run out of conversion capacity.
  2. Peers and advisers. A Cornerstone Advisors partner, quoted by PCBB, recommends asking colleagues at other institutions how they feel about their core provider, “and not just those who offered testimonials.” Cornerstone Advisors (opens in a new tab) itself says it has run more than 700 core evaluations.
  3. Request for proposal and negotiation. Formal proposals, demonstrations, reference calls and contract terms follow.
  4. Due diligence. US regulators expect banks to manage the risks of every third-party relationship. In 2024 the Federal Reserve, FDIC and OCC issued a guide for community banks (opens in a new tab) that walks through “each stage of the third-party relationship,” building on the agencies’ June 2023 guidance.
  5. Conversion. Industry advisers quoted by PCBB say it can take one to three years to select a new core and convert, so institutions should start 24 to 30 months before the contract expires.

Fraud and payments decisions move faster than core decisions but follow the same logic. In Alloy’s 2026 fraud report (opens in a new tab), a survey of banks, credit unions and fintechs by a fraud-prevention vendor, 67% of respondents saw an increase in fraud attempts over the past year, and 71% said fraud most commonly occurred in online or mobile banking channels. Instant payments add urgency: the FedNow Service (opens in a new tab) settled 4,997,811 payments in the second quarter of 2026, with volume up 83.2% on the previous quarter.

Where do AI assistants enter that buying process?

Mostly at the early market scan, the step before any vendor knows a bank is looking.

No public survey measures how bank technology buyers use AI assistants, so the evidence comes from wider business buying:

  • In a buyer survey by Responsive (opens in a new tab), a company that sells proposal software, 48% of US buyers said they use generative AI for vendor discovery. In Digital Commerce 360’s coverage (opens in a new tab) of the same research, 90% of buyers said they research before first contact, and the weight placed on industry expertise was “especially strong in technology and financial services.”
  • The TrustRadius 2026 report (opens in a new tab), as summarized by MarketScale, found 63% of technology buyers used AI tools during their purchase journey, and the average shortlist held just 2.7 products.
  • Banks are building their own AI habits. In Bank Director’s survey, 66% of banks had drafted an acceptable use policy for AI and 62% were experimenting with it in limited use cases.

Put together, a reasonable expectation is this: when a COO is asked to “see who else is out there” two years before renewal, an AI assistant is now one of the first places that question goes, alongside peer calls and a consultant. And with short shortlists, being absent from that first map is costly. Many professional buyers start with a favorite: 61% of Responsive’s respondents said they begin with a preferred vendor in mind, though 45% said they are open to switching.

Which questions do bank buyers ask an assistant about technology vendors?

Questions about fit by institution size, conversion risk, integration, fraud coverage and regulatory readiness. The examples below are illustrative, written by us; they are not observed prompts.

StageIllustrative prompt
Market scan“Which core banking providers serve credit unions between $1 billion and $3 billion in assets?”
Comparison“How do cloud-native cores compare with traditional providers for a community bank?”
Conversion risk“How long does a core conversion take, and what usually goes wrong?”
Payments“Which providers help a community bank receive instant payments through FedNow?”
Fraud“Fraud platforms that cover check fraud and real-time payment scams for regional banks”
Due diligence“What should we ask a core vendor about resilience and financial stability before signing?”
Alternatives“Alternatives to our current core if we only want a sidecar for digital accounts”

Two things stand out. First, many of these questions are about risk, not features, because a failed conversion is a board-level event; one industry association executive calls a core change “like open-heart surgery.” Second, several questions are really about capacity and timing, which few vendors state publicly.

How does an AI answer turn into a signed bank contract?

By putting the vendor on the list that becomes the RFP invitation list. The rest is the vendor’s own sales work.

The path, as we see it:

  1. Named in the market scan. An assistant, a peer or a consultant names three or four vendors.
  2. Checked. Buyers verify what they read. In the TrustRadius report, 94% of buyers who used AI said they fact-check its outputs at least some of the time. For a bank, that means vendor pages, trade press, peer calls and, increasingly, a second AI question.
  3. Invited to the RFP. In Responsive’s survey, the proposal response was the most important factor shaping the decision, cited by 81% of buyers, and industry expertise carried significant weight for 52%.
  4. Due diligence and contract. Third-party risk review, negotiation and a multi-year agreement.
  5. Expansion. Once the core is live, the vendor sells payments, card and digital services on top. Jack Henry’s results show the effect: full-year processing revenue rose 8.2%, faster than total revenue, which grew 7.1%.

The AI answer sits at step one, which is why its value is easy to miss. A vendor sees an RFP arrive and credits the relationship; the bank’s team may have first met the vendor’s name in an assistant two years earlier. We suggest asking prospects, at the first meeting, where they first heard of you, and logging which RFP invitations came without a prior relationship. More on that measurement gap is in why analytics miss AI visibility.

What decides whether a banking technology vendor is named?

Clear facts and independent evidence; platforms document how they search, not how they choose vendors.

Documented by the platforms. Google says AI Overviews and AI Mode may use “query fan-out” (opens in a new tab), issuing multiple related searches before answering. OpenAI says ChatGPT search (opens in a new tab) “typically rewrites your query into one or more targeted queries” that it sends to search providers. Neither publishes how it ranks banking technology vendors.

Observed in our studies.

  • Business software questions almost always get an AI answer. In our frequency study, 96.0% of US B2B software and technology keywords triggered an AI Overview.
  • They draw less on page-one results than finance does. In our citation study, 23.9% of B2B software citations were page-one Google results, so pages beyond the top ten, such as trade coverage and documentation, get cited too.
  • Named sources tend to be cited. In our hidden-searches study, when ChatGPT’s own search named a source, the answer cited it 44.0% of the time, against 8.1% when it did not.
  • Self-promotion is visible. In our study of “best” lists cited by AI, 24.2% of numbered lists with an identifiable publisher ranked their own publisher first.

Our inference, specific to banking technology. The facts a bank needs most are the ones vendors often keep behind a sales call: which institutions of what size run the platform, how long conversions take and how many are booked, which payment rails and fraud types are covered, which independent audits exist, and how the company is funded. An assistant can only repeat what it can find and check. Trade association surveys, banking press, consultants’ public reports and customer announcements are the kind of independent sources we expect an assistant to rely on, and the kind a risk-minded buyer trusts.

What does GEO look like for a vendor selling to banks and credit unions?

Generative engine optimization (GEO) makes your fit, track record and risk facts easy for assistants to find and repeat.

For a banking technology vendor, the work usually covers:

  1. A clear entity. One plain description of what the platform is (full core, sidecar core, payments hub, fraud platform), which institutions it serves by type and asset size, and where it runs.
  2. Named proof. Public customer announcements with permission, conversion case studies with dates, and the institutions’ own statements, so “who uses it” has a checkable answer.
  3. Due-diligence facts in public. Resilience, security audits, data handling and ownership, written for the third-party risk questions banks must ask, reviewed by your legal team.
  4. Independent coverage. Banking trade press, association research, conference sessions and consultant briefings, which an assistant can cite and a buyer can verify.
  5. Fair comparison pages. Honest pages on how you differ from incumbents and from cloud-native challengers; our comparison-page guide covers what the evidence says.
  6. Monitoring. Ask assistants the questions your buyers ask each quarter and correct wrong facts at their source, following how to fix wrong information about your brand in AI answers.

None of this guarantees a mention; no one can promise how an assistant will answer. GEO makes sure that when a bank’s team asks, the evidence about you is complete, current and accurate. How enterprise shortlists form more broadly is covered in are AI assistants shaping which enterprise software gets shortlisted?.

What is still unknown about AI and bank technology buying?

Plenty: there is no public measure of how often bank buyers use AI, or of RFP invitations it produces.

  • No bank-specific data. The AI-use figures above come from wider business buyer surveys, two of them run by vendors (Responsive sells proposal software, Alloy sells fraud prevention). We found no survey of bank technology buyers’ AI use.
  • Private assistants are invisible. Many banks use assistants inside their own controlled tools. What those tools search, and which sources they can reach, is not public.
  • Our studies are snapshots. They cover US searches collected in September 2026; answers change over time and between assistants.
  • The link to revenue is unmeasured. With conversions this rare, no vendor has published how many RFP invitations trace back to AI answers. Our review of whether AI visibility drives business results sums up what other industries have measured so far.

Where should a banking technology vendor start?

Start with the questions banks ask two to three years before renewal, and check what assistants say about you.

A useful first review covers the early market-scan questions in your category, the risk and conversion questions buyers ask about you by name, and the comparisons with incumbents. It shows whether you are named, which competitors and sources appear instead, and whether your customer base, conversion record and due-diligence facts are described correctly.

If your growth depends on being invited into core, payments or fraud evaluations before the RFP is written, talk to us about a review of your standing in AI answers. We will compare how assistants describe you and your competitors for bank and credit union buyers, list what they miss or get wrong, and plan the content, coverage and proof that can earn you a place on the next shortlist. Our generative engine optimization service page details how we put named proof, conversion records and due-diligence facts where bank and credit union buyers, and their assistants, can check them.

Frequently asked questions

Do bank executives really use ChatGPT to research core vendors?

No public survey measures it for banks. Wider surveys suggest many do: 48% of US business buyers in Responsive’s research use generative AI for vendor discovery, and most banks are now writing AI use policies.

Why does AI visibility matter if banks buy through consultants and RFPs?

Because the shortlist forms first. Consultants and RFPs decide among the names already in view, and the average shortlist in the TrustRadius report held only 2.7 products.

Can a newer cloud core compete with incumbents in AI answers?

It can be named if the evidence exists: live institutions, conversion records and independent coverage. Assistants cannot recommend what they cannot verify.

Should we publish our conversion capacity and timelines?

Where your business allows, yes. Banks ask about timing early, CUInsight reports slots booking into 2029, and a vendor that states its capacity gives assistants and buyers a fact to repeat.

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