Case Studies/US Based Insurance Payment Provider
How a twelve-person fintech earned the citation share of a team three times its size.
Ninety days. Four times the AI citations. The mechanism, described.
This company is an insurance payment provider with the disposition of a small atelier. They had built a product their customers genuinely loved. The audience that should have known it, finance leaders at mid-market companies running cross-border payouts, was instead being told about competitors when they asked the four models. The work was excellent. The presence was incomplete.
AI citation share of voice across the four models, measured at 95% Wilson confidence.
Category-defining queries lifted from no rank to position three on average.
LinkedIn organic reach within ninety days of activation, without paid amplification.
i / What We Found
A 17% citation share. Strong on transactional queries. Absent on category-defining ones.
The first scan returned a 17% citation share across the four models, meaningful but proportionate to a much smaller company. Search positions were strong on transactional queries; their existing SEO discipline was sound. They were absent on category-defining queries, the questions a buyer asks before they know which vendors to evaluate. Social signal was concentrated on LinkedIn, dispersed elsewhere. The orchestration layer was missing entirely.
The diagnosis was specific. The engine could see what the team had built, what was working, and what was unconnected. The remediation plan was not a list of tactics; it was a sequence of motions, ordered by which would compound first.
ii / What We Built
Thirty-eight articles. One hundred and forty-two technical fixes. The orchestration layer activated.
Quill produced thirty-eight articles in ninety days, each written in one brand voice and structured for AI extraction across search, social, and the four models. Server-side fixes resolved one hundred and forty-two technical debts the previous SEO instrument had identified but never written. Schema deployed across the entire content tree. The flms.re protocol activated.
Social Signal listened across the five layers. The system drafted eighty-four responses in the US based insurance payment provider's AI voice; the human team published seventy-one of them. Brand voice was tuned by ingesting the existing public communications corpus, the Slack-shaped warmth that distinguished their tone from a category that defaults to formal financial language.
The orchestration layer connected the work to the loop. Citations from AI engines fed the social calendar; engagement on Reddit threads fed the next article; rank movement fed the next Quill brief.
iii / What The Engine Produced
The loop closed on day 60.
The compounding has not stopped.
AI citation share rose from 17% to 72% across the four models, measured at 95% Wilson confidence. Category-defining queries moved from no presence to position three average. LinkedIn organic reach quadrupled, without paid amplification, without an outbound campaign, without a product release.
The loop closed in week eight. By day sixty, content cited in AI answers was being shared on LinkedIn, which was building topical authority, which was lifting search rank, which was being cited in AI answers. The engine ran on its own velocity.
By day ninety, the second wave was visible. Adjacent topics, those the team had not yet briefed, began to rank because the model had learned the brand on the first set. The compounding was no longer linear; it was geometric.
"We did not change our work. We changed who could find it. CLEO carried the difference."
COO