Introducing Cite

An AI revenue agent that works the new search surface.

Mersel is an enterprise AI agent for AI search. This is its command center: brand mentions, share of voice, and citation rate tracked across every model — ChatGPT, Perplexity, Gemini — for one brand. I started the company and lead product and design from the ground up.

The hardest design problem was authoring. The agent researches a brand and writes content structured so an AI engine can parse and quote it — at volume, twenty-eight of a hundred pages here. The work is autonomous, but a human still has to steer it and trust it, so I built the surface around that progress and review.

Once written, the agent publishes the pages to the brand's own site as a resource center — on their domain, in their design. The decision here was to make the agent's output look native: not a third-party widget, but pages a company would be proud to host.

It doesn't stop at publish. The agent keeps the pages current as the brand and the models shift — updating, re-citing, re-ranking. Designing for an always-running agent meant showing that work is happening without turning the interface into noise.

AI visibility only matters if it converts. The funnel ties the whole loop back to revenue — visitors arriving from AI-cited pages, and the leads they become. The design job was to make a channel no one can watch directly read as concrete pipeline a sales team can act on.

This is the proof. Each lead carries its page journey — the exact AI-cited page that brought them in, from first landing to submission. Closing that attribution loop is what makes an invisible channel believable: you can point at the citation that produced the deal.

The loop ends as named, verified companies handed to sales — real buyers, not anonymous traffic. I shaped the product narrative end to end, designed every surface, and built the brand and marketing site so the whole thing reads as one system: an agent that owns a channel, from the first audit to the last citation.