Composite Company Profile
A US-headquartered enterprise fintech platform selling treasury and payments software into mid-market and enterprise finance teams. Series D, roughly $80M ARR, $180K-$450K ACV, sales-led motion with a mature outbound org and a 4-6 month sales cycle. The buying committee typically included a VP Treasury, a CFO signer, and a Head of Payments Ops, with security review from an internal InfoSec team.
The Problem
The client had spent two years building a strong SEO footprint on classical Google search. They ranked in the top three for a cluster of high-intent commercial terms tied to treasury automation, cash forecasting, and B2B payments orchestration. Organic pipeline was healthy on paper. Then two things started happening in parallel.
First, branded and unbranded impressions in Google Search Console flattened, then dipped roughly 14% quarter over quarter, even as their published content velocity increased. Second, their AEs began reporting that new pipeline prospects were entering discovery calls already referencing competitor product features, integration lists, and pricing anchors that the client had never volunteered. When AEs asked how the prospect had reached a shortlist, the answer was increasingly the same: ChatGPT, Perplexity, or an internal Copilot deployment surfacing a competitor as the recommended treasury vendor.
Prior efforts had focused on more blog posts, more backlinks, and a Rank Math schema audit. None of that moved the needle because none of it addressed the actual channel where the buying committee was now doing early vendor screening.
What Our Team Diagnosed
Our SEO and analytics team ran a 30-prompt AI citation audit across ChatGPT, Perplexity, Gemini, and Claude, plus Google AI Overviews, targeting the exact discovery-stage queries the client’s ICP was asking. The result was blunt. The client was cited in 8 of 30 prompts. Two direct competitors, both smaller by ARR, were cited in 22 and 19 prompts respectively. The client’s own content was structurally hostile to AI retrieval: long marketing preambles before the answer, PDF-gated data, missing entity schema on the company node, no llms.txt file, and product pages that never named the underlying job to be done in a way an LLM could extract as a passage.
Separately, we discovered that a large share of their strongest ranking pages were 2,400+ word thought-leadership posts with the actual product-fit answer buried in the middle. LLMs were finding the competitors because the competitors had shipped short, entity-dense, quotable answers on top-of-funnel pages that read like reference material.
Strategy MV3 Shipped
We proposed a Generative Engine Optimization (GEO) engagement layered onto the existing SEO program, not a replacement. The remit was to make the client the default citation on the top 60 discovery prompts their ICP was asking generative engines, without cannibalizing their existing classical SEO rankings.
Key strategic choices:
- Treat AI citation share as a first-class KPI alongside impressions, clicks, and pipeline. Instrument it weekly.
- Rewrite the top 22 commercial pages with a Quick Answer block in the first 60 words, followed by entity-dense supporting evidence, then narrative depth.
- Ship a full schema pass: Organization with sameAs, Product, FAQPage, HowTo where relevant, and BreadcrumbList site-wide.
- Publish llms.txt, expose a machine-readable pricing summary page (behind a soft gate, not a PDF), and de-gate the top four whitepapers into HTML with schema.
- Build out a comparison hub answering the exact vendor-versus-vendor prompts the buying committee was pasting into ChatGPT.
Implementation
The engagement ran under the Scale AI tier. Vance oversaw strategy and QA. Our SEO team led the GEO rewrite, our analytics team stood up the citation tracking dashboard, and our content team executed the comparison hub and Quick Answer rewrites.
Deliverables shipped over the first 90 days:
- 22 commercial pages rewritten with Quick Answer blocks and entity schema.
- 14 new comparison and category pages built for prompts the ICP was actually asking generative engines.
- llms.txt deployed and referenced from robots.txt and the sitemap index.
- Weekly AI citation tracking across ChatGPT, Perplexity, Gemini, Claude, and Google AI Overviews for a locked 60-prompt basket.
- Migration of four gated PDFs to HTML with proper Article schema and TL;DR summaries.
- A 6-touch nurture sequence for prospects entering discovery who had first interacted via an LLM referral, tracked in the CRM as a distinct source.
Cadence was weekly stand-up with the client’s Head of Growth, a monthly steering review with the CMO, and a fortnightly citation-share readout to the demand gen team.
Outcomes
Measured at day 180 against a locked pre-engagement baseline:
- AI citation share on the 60-prompt basket moved from 27% to 71%, overtaking both primary competitors.
- Google organic clicks recovered from the 14% dip and grew 22% net above the original pre-decline baseline, driven partly by AI Overview inclusion sending qualified click traffic.
- Self-reported “found you via ChatGPT or Perplexity” on discovery call forms went from 4% to 31% of net-new pipeline meetings.
- Sales-qualified pipeline attributed to organic and AI-referral sources grew 2.4x over the two prior comparable quarters.
- CAC on the organic + AI segment fell 38% versus the trailing paid-social and outbound blended CAC.
The primary outcome the client cared about, and the metric we anchor this case study on, is the 2.4x lift in sales-qualified pipeline from organic and AI-referral combined.
Timeline
Kickoff to first measurable citation-share improvement: 6 weeks. Kickoff to full outcome measurement: 6 months. The engagement continues on a Scale AI retainer with a rolling 60-prompt citation basket refreshed quarterly.
NDA Framing
Client identity is protected under NDA. Industry, ARR band, ACV range, and outcome metrics above are accurate to the engagement. Company name, brand assets, and direct quotations are withheld. Full details, including the citation-tracking methodology and the pre and post prompt-by-prompt results, are available under mutual sign-off during a discovery call.
Composite Testimonial
“We had rebuilt our SEO program twice in three years and were still watching competitors get cited over us in every LLM our prospects were using. MV3 was the first team that treated AI citation as a measurable channel with its own playbook. Six months in, the pipeline math moved.” — Priya, VP Demand Generation
Next Step
If your buying committee is running vendor shortlists through ChatGPT, Perplexity, or Copilot before your SDR ever gets a reply, and you cannot tell whether you are in the citation set, that is the problem we solve. Book a discovery call or review our GEO Audit entry offer to see where you stand on the prompts your ICP is actually asking.
