Type of SaaS
Financial (FinTech)
SERVICES PROVIDED
Generative Engine Optimization (GEO), SEO, Content Marketing
Results Window
May–September 2026, based on daily tracking of 25 commercial category queries across ChatGPT, Gemini, and Perplexity.
The Situation
Buyers in this market research heavily before they talk to anyone, and that research now starts in AI assistants. When a firm asks ChatGPT, Gemini, or Perplexity to recommend software for its practice, the answer shapes the shortlist before any vendor hears the question. Missing from those answers is invisible lost demand.
In May 2026 the client was present but losing. It ranked second in its category with 1,090 citations across the month, yet the daily tracking showed a brand consistently mentioned rather than consistently recommended, usually around position three in the answers that mattered. Category incumbents dominated the citations, and their listicle pages did the heavy lifting. Competitor “best software” roundups were being fed into AI answers again and again, while the client’s most-cited page, its home page, earned 302 citations.
The client needed to move from “also mentioned” to “the source the engines quote first,” and it needed a program that could prove, month over month, whether that was happening.
Key Results
| METRIC | BEFORE | RESULT |
|---|---|---|
| Monthly citations in AI answers, all pages | 1,090 (May 2026) | 3,570 (September 2026) |
| Most-cited owned page, monthly citations | Home page: 302 (May 2026) | Category listicle: 754 (September 2026) |
| Brand visibility in tracked AI answers | 64.83% (July 2026) | 86.33% (September 2026) |
| Rank among most-cited domains in the category | #2 (August 2026) | #1 (September 2026) |
| Monthly citation gap vs. the category incumbent | Trailing by 140 (August 2026) | Leading by 1,616 (September 2026) |
What We Did
Audited every AI answer to find out where citations actually come from.
We tracked a fixed panel of 25 commercial category queries daily across ChatGPT, Gemini, and Perplexity and audited every source cited in the answers. The audit produced the insight that drove the program: citations in this category were overwhelmingly sourced from vendor websites, and among vendors, listicles dominated. Competitor roundup pages were winning the recommendation slots while the client's pages were mentioned as context.
Published the definitive category listicle, engineered for AI retrieval.
The category's core "best software" questions were being answered with competitor roundups, so we built the client's own definitive listicle for the category's top query, structured so an AI engine could extract, verify, and cite it. By September the listicle was earning 754 citations a month, more than double the home page's 302 in May, on a page that did not exist when the window opened.
Refreshed the pages AI engines already trusted.
Citation wins compound when the pages engines favor stay current. We refreshed copy and consolidated internal links on the client's most-cited feature and comparison pages, including the flagship feature page, among the most-cited pages on the site throughout the window, and the enterprise-segment pages, which began appearing earlier in answers for large-volume and scaling workflows. The refreshed pages are the ones showing sustained month-over-month citation growth. The pages we did not refresh held flat.
Strengthened the entity footprint where AI engines look for corroboration.
AI engines corroborate what a site says against the wider web. We tightened the client's entity presence across association partner pages, directories, and third-party review profiles so independent sources consistently confirmed the brand's position and claims. By September, the client's review profiles and association partnerships were showing up as cited sources in AI answers, adding off-site authority the client did not control before.
Wired the CRM so AI-sourced demand could be measured, not assumed.
We tagged lead source in the client's customer relationship management (CRM) platform so contacts arriving from ChatGPT, Perplexity, and Gemini are identified. The data is early but real: 14 AI-sourced contacts were recorded in 2026 through October, versus 13 in all of 2025, including the first contacts attributed to Gemini. Citation wins precede demand. Now the client can measure that lag instead of asserting it.
What The Data Tells Us
Citations more than tripled, from 1,090 in May to 3,570 in September, while the number of cited pages barely moved, from 43 to 45. Average citations per cited page rose from roughly 25 to roughly 79. Growth did not come from publishing more pages. It came from two moves working together: the listicle added a new page that engines could quote for the category’s biggest question, and the refreshes pushed the pages engines already trusted up the citation rankings. The sustained increases concentrated on the refreshed pages, which is the clearest signal in the data that the program drove the change rather than an engine update. It also validates the core generative engine optimization (GEO) thesis for this category: the “best software” questions were being answered with listicles, and the client now owns its category’s definitive one instead of a competitor.
The competitive reversal confirms the strategy held. In August the client trailed the category incumbent by 140 citations. In September it led by 1,616, a swing of more than 1,750 citations in a single month, while brand visibility in tracked answers rose from 64.83% in July to 86.33% in September. The remaining gap, roughly 14% of tracked prompts, is now the program’s measurable target rather than a mystery.
Bay Leaf Digital on Generative Engine Optimization (GEO)
We build GEO programs for B2B SaaS companies that make you the source AI engines cite when buyers ask about your category, before your traffic data shows the conversation has moved.