Building an AEO Program from Scratch: 40% More LLM Citations on a Priority Query Set

This engagement built an AdTech SaaS company’s first integrated SEO and AEO program from the ground up—increasing organic traffic by more than 30% in six months and lifting LLM citation rates 40% across a stable priority query set. Here is how it worked, and what I would refine next time.

The Challenge

An AdTech SaaS company needed organic growth in a competitive space. But beyond Google rankings, there was a larger problem: most brands aren’t optimizing for how AI actually cites content. ChatGPT, Perplexity, and Google SGE were reshaping discoverability with no existing playbook.

The Strategy

Built the company’s first integrated SEO and AEO program from scratch. Built a monitoring framework tracking content surfacing across ChatGPT, Perplexity, and Google SGE. Tested entity schema, published 25+ long-form articles, implemented FAQ schema, and built GA4 attribution aligned to go-to-market priorities.

The Results

30%+ organic growth within six months. Lifted LLM citation rate 40% across a stable priority query set in ~3.5 months through systematic schema and entity optimization. Built the AEO measurement framework in Profound, plus the content playbooks the program still runs on.

What This Taught Me

The future of SEO is about making your content the most discoverable across all search paradigms—traditional engines, AI answer systems, and human trust. Organizations that understand this now will dominate visibility later.

What I’d Do Differently

I’d establish a longer pre-launch citation baseline. We began monitoring LLM citations around the same time content production started, so the 40% increase is real, but the attribution is noisier than I’d prefer. Sixty to ninety days of baseline data would have created a cleaner before-and-after comparison and made the results more defensible.

The competitor citation analysis also needed to be more systematic. I compared five to seven cited pages around each pillar, using Claude to identify shared coverage and missing information, then researched and filled the gaps independently. The missing step was reverse-engineering each winning page on its own: how it framed the answer, structured its evidence, connected relevant entities, and made its claims easy to extract.

That analysis would probably have pushed us toward fewer, deeper pieces earlier. Instead, we learned through a full production cycle that citation gains were concentrated in a smaller group of highly structured, entity-anchored articles.