AI Referrals Are Surging; Are You Ready for the New Customer Journey?
By Nicole Volpe, Contributor at The Financial Brand
Simple Subscribe
Subscribe Now!
Executive Summary
- Traditional organic traffic is declining 15-64%, but AI-referred visitors convert at 3x higher rates because they arrive further down the purchase funnel with decisions already made.
- Financial institutions must optimize for AI engines, not just search engines – This requires structured, machine-readable content across multiple channels.
- AI now acts as an “outside sales channel,” guiding consumers through research and comparison before sending them to banks ready to purchase, demanding institutions prepare for seamless, streamlined conversion processes.
Financial institution marketing teams are well past wondering if the slow-rolling web performance declines they’ve been observing are AI’s fault. Instead, they’re asking harder questions: How do we get our traffic back? And more pointedly: How do we make sure our institution’s products and services are showing up in AI-generated responses?
Since going mainstream about 30 months ago, AI has had a two-pronged impact on consumer search and discovery. First, zero-click AI-generated results within traditional search engines are significantly cutting into organic traffic (by 15% to 64%, according to one study). At the same time, the portion of referral traffic from platforms like ChatGPT, though still quite small, has been surging.
For banks and credit unions that depend on predictable flows of organic and paid traffic for new customer acquisition, the first dynamic feels like unmitigated bad news: the beginning of the end of a vital discovery channel. What they may not immediately see, however, is that the second impact — growth in referrals from platforms like ChatGPT, Perplexity, and Gemini — is an opportunity with distinctive upsides for institutions that adjust their best practices.
Milestone — an AI-first digital experience platform with a focus on financial services and other industries — has analyzed data from roughly 3,500 client sites. Its research underscores what many marketers have been experiencing: AI-driven traffic to its sites has climbed more than 500% since January 2025. Today this represents only about 0.3% of total traffic on the sites Milestone tracks, but by 2026 Milestone projects AI’s share will rise to 27%. Meanwhile, the company projects that traditional search’s share of total site traffic will fall from 55% to 27%.
Milestone’s projections suggest a growth ramp similar to that of the mobile explosion 15 years ago. And marketers need to reckon with this transformation as surely as they did with its predecessor.
The New Journey
Of course, as emerging AI users themselves (isn’t everybody?) most marketers and consumers are already familiar with the new journey’s basic outlines. People who turn to AI aren’t looking for the deep lists of links they might have gotten from Google or Bing. What these users want are direct, context-rich answers, delivered in output formats that are specifically adapted to the problem they’re trying to solve.
In the view of Milestone CEO Anil Aggarwal, that’s where the opportunity lies: Conversions from AI-referred visits are roughly three times higher than those from traditional search, according to Milestone’s data. That’s because they’re further along in their decision process; they’ve filtered options, compared features, and narrowed their choices.
“The biggest shift now is that the AI engines are giving you the answer and the links they provide are secondary,” Aggarwal said. “Previously, Google handed off the traffic, leaving the decision to be made on an institution’s site; now, traffic doesn’t flow out until the very last moment, when the purchase decision has already been made.”
But, said Aggarwal, “You don’t need to panic, you just need to plan.” For banks and credit unions, that means ensuring your institution is in position to benefit from this new, higher-value customer journey.
Everyone’s ‘Feeling Lucky’ These Days
As familiar as marketers may be with the new search experience, many know less about the underlying model differences that are driving the new result outcomes.
Traditional search engines return results by scanning their index for pages that match a query, then ranking those pages based on signals such as keywords, backlinks, and click history. AI search engines, on the other hand, analyze the intent behind a query, draw from a wider range of indexed sources, and cross-check those sources against each other. Rather than prioritizing the most popular pages, they emphasize information that is consistent across multiple sources, backed by authoritative references, and up to date.
This extra layer of cross-checking and validation distinguishes AI search from traditional search. It means financial institutions have to make sure that their information is available across multiple channels and is also consistent and exhaustive. This makes the task of marketing more complex but leads to better consumer experience.
The AI-driven workflow shifts consumer research from gathering individual pieces of information across multiple institution’s websites to receiving synthesized explanations that highlight trade-offs and context. The consumer sees information organized into cohesive narratives: for example, credit cards explained in terms of fees versus perks or mortgage options framed around monthly payments versus long-term cost.
Some users may try to exert more direct control over the result, asking for comparison charts mapped to specific criteria or top ten lists or even top three lists. Many others are happy to let AI infer the best presentation format based on their intent. Either way, the research process gains depth and coherence, feeling more like guided consultation than traditional search. As a result, AI search pushes consumers farther down the funnel, functioning almost like an outside sales channel. That makes it essential for AI agents to surface the strongest, most accurate information about your offerings, because they, not you, are the ones guiding consumers to the decision point.
Dig deeper:
- How the Agentic AI Revolution is Transforming Operations at 70% of Banks
- Four AI Action Plans to Supercharge Your Marketing Team’s Impact
- From KYC to KYAI: Why ‘Algorithmic Transparency’ is Now Critical in Banking
Critically, to appear in AI results, the financial institution’s product content must include clear, machine-readable eligibility criteria and dynamically updated attributes, such as minimum balances, fee structures, and insurance coverage. Someone comparing credit cards may toggle between travel rewards, cash back, or balance transfers, seeing trade-offs update with each shift in parameters.
Aggarwal described a recent case in which a bank’s offering failed to appear in the response to an AI query for the best CD rates. ChatGPT had added a requirement that the product be FDIC insured. Although the bank did meet that standard, the disclosure appeared only in an image on its website that was unreadable to the AI system; and the institution was left out of the results.
This new consumer-led process of iterative refinement and scenario testing demands greater precision from providers at each step in the process. By the time the user reaches a bank’s website ready to apply, they expect every detail to match what AI has already summarized, placing yet more pressure on institutions to be accurate and consistent.
Adapt to Transform
For financial institutions, meeting customers as they progressively grow into this new journey means leveling up their own digital content strategy. Aggarwal identifies five pillars for banks and credit unions to focus effort on.
1. Augment SEO with GEO. Traditional SEO remains necessary but is insufficient. Generative Engine Optimization (GEO) demands more. Institutions need to publish information in structured, machine-readable formats so that FAQs, rates, and other product details can be interpreted accurately and pulled directly into AI-synthesized answers. Getting this right requires understanding how AI engines derive their answers. In the example above, in the search for “the best CD rates”, the result is based on rate competitiveness, liquidity and term variety, FDIC/NCUA insurance, institution type, freshness of data, and consensus. (The latter refers to whether multiple rate aggregators validate the source.)
2. Make your website as data hub. Aggarwal emphasized the importance of establishing your website as a centralized “data hub” that serves as an authoritative repository for consistent, up-to-date product content. Rather than scattering information across multiple pages or formats, institutions should consolidate details (e.g., eligibility criteria, rate updates, and compliance language) into a single, easily accessible source. This structure supports consistency across both human-facing websites and AI-driven responses.
3. Invest in conversational agents. Customers accustomed to the dynamic Q-and-A of ChatGPT’s user interface will come to expect a similar experience on a bank’s website. Moving beyond rules-based chatbots toward AI-powered agents makes it possible to answer nuanced questions, maintain brand voice, and hand off seamlessly to human staff when needed.
4. Enable personalization. If financial institutions haven’t yet absorbed the lessons of Netflix, Amazon, and a host of fintech competitors, now is the time to step up. Increasing time spent with AI is raising consumer expectations that their bank and credit union interactions likewise reflect their personal choices, past purchases, query history — tailoring content and offers to individual likes and dislikes, and intent, and dynamically adjusting digital experiences to what customers have already signaled they want. According to a study by McKinsey: 71% of consumers expect personalized experiences and 76% of consumers get irritated with generic, nonpersonalized content. (The study also shows that personalized marketing actions or tactics yield 40% more revenue.)
5. Streamline conversion. As AI tools evolve toward enabling direct completion of tasks, such as starting an account or loan application inside the engine itself, banks and credit unions must prepare their systems for seamless handoffs. Each of these purchase-enabling, last-mile tools should be easy to use for humans and the AI Engines. Once a customer is ready to act, the process should be streamlined and error-free. The key is to minimize steps between research and action, whether for humans or AI.
At first glance, these pillars may seem like an overwhelming volume of change for a financial institution to manage. But in reality, all of them align with challenges institutions are already addressing. GEO, for example, naturally evolves from SEO. And building personalized experiences has long been a priority due to competition from fintechs and digital-first banks. While AI demands new expertise, it also offers the promise that financial institutions, especially smaller ones, can achieve parity by engaging directly with consumers.
