How a B2B SaaS Agency Proved AI-Driven Revenue to Three Clients in One Quarter

An agency running AEO for three B2B SaaS clients walked into three separate QBRs last quarter with the same slide: a dollar figure, per engine, per post, tied to closed deals. Not sessions. Not rank. Revenue. That is the shift Clarity's attribution layer makes possible for portfolio work, and the three-client picture is more interesting than most people expect, because the engines do not behave the same way twice.

Here is what the numbers actually showed, and why the mix mattered more than the totals.

The setup: three clients, one 90-day window

The agency, a five-person shop working with early-growth B2B SaaS, connected each client's site and analytics to Clarity in the first week of the quarter. Clark ran the site scan and the brand-voice interview for each account, Prompt Monitoring started asking buyer questions across ChatGPT, Claude, Gemini, and Perplexity daily, and Clark began publishing answer-first posts on a weekly cadence to each client's existing CMS.

The three clients were deliberately different:

• Client A: a workflow automation platform selling to mid-market ops leaders, $18K average deal, 45-day cycle.
• Client B: a compliance tool for fintech startups, $6K average deal, 21-day cycle.
• Client C: a data infrastructure product for engineering teams, $60K average deal, 90-day cycle.

Same agency, same playbook, three very different buyers.

What the attribution data actually showed

By day 90, the agency had per-engine, per-post attribution for each account. The totals were useful. The distribution was the story.

• Client A (workflow automation): ChatGPT drove 62% of AI-attributed leads. Two how-to posts Clark wrote about a specific integration workflow accounted for the bulk of it. Board-ready number: seven closed deals traced to ChatGPT citations, one of which was a $34K expansion.
• Client B (fintech compliance): Perplexity punched above its weight. It sent fewer sessions than ChatGPT but converted at nearly 3x the rate, because Perplexity users arrived deeper in the evaluation. Claude was a close second. ChatGPT was third. The mix was almost the inverse of Client A.
• Client C (data infrastructure): Gemini and Claude split the pipeline roughly evenly. The 90-day cycle meant only two deals closed in-window, but the pipeline value tied to AI-cited posts was north of $180K, and both closed-won deals came from a single deep-dive post about a niche architecture question that competitors were not answering.

Three clients, three different engine profiles, three different content shapes doing the work. If the agency had assumed ChatGPT would dominate all three, they would have been wrong twice.

Why standard reporting could not have produced this

Before Clarity, the same agency was showing clients ranking movement and traffic charts, then handing over CRM exports and asking clients to trust that the two were related. Most AI sessions were landing as direct or unattributed traffic. The engines strip referrers. Standard analytics cannot tell you a ChatGPT citation turned into a lead, let alone which post got quoted.

Clarity's attribution scheduler writes a per-business, per-day, per-page, per-source snapshot that ties the session through to the conversion and the dollar amount, and flags whether the landing page was one Clark wrote. That is the loop. Content, monitoring, and attribution talking to each other, once a day, per client, with a stamp on the pages Clark authored so the agency can prove the work is what earned the deal.

For the agency's QBRs, that turned into three clean statements:

• "ChatGPT drove $126K in closed revenue this quarter, from these two posts."
• "Perplexity converted at 3x the rate of any other engine; here are the four posts driving it."
• "Gemini and Claude are your enterprise pipeline; here is the $180K they are sitting on and when it should close."

No client asked whether the numbers were real. They asked what to publish next.

The takeaway for agency owners

If you run AEO across a portfolio, the engines will not behave uniformly, and that variance is the argument. A three-client sample is enough to show a board that AI visibility spend is a channel with per-engine economics, not a single bet. What you need is the loop: monitoring that finds the gap, content that fills it, and attribution that traces the dollar back to the exact engine and post.

Clarity's agency pricing is built for this, with graduated multi-site billing that scales with the portfolio. If you are trying to walk into your next QBR with a revenue number instead of a rank chart, start with the free diagnostic on one client site and see the visibility gap before you decide anything else.