Twelve months ago, generative engine optimization was a theory my team was actively discussing but hadn’t fully committed to for any client campaign. We understood the concept. We’d seen the early research. But we were cautious about recommending a significant strategic shift based on a trend that was still developing.

Then one of our longer-standing clients, a mid-size B2B technology company with a healthy organic traffic baseline, agreed to run an experiment. Six months of deliberate GEO optimization alongside their existing SEO program, with careful before-and-after measurement.

Here’s what we found.

The Setup

The client’s situation was reasonably representative. Strong domain authority, solid technical SEO foundation, content that ranked well for target keywords but had been built primarily around traditional search optimization. No particular attention had been paid to how their content was being used or cited in AI-generated search responses.

We ran a baseline measurement period first. Tracked which of their target queries were triggering AI Overviews in Google. Manually audited which sources were being cited in those overviews. Analyzed the characteristics of the content being selected versus the client’s content covering the same topics.

The pattern was consistent. The content being cited in AI Overviews tended to be more specifically structured, with clearer direct answers to the implicit question behind each query. It tended to use more concrete examples and specific data points. It tended to have cleaner information hierarchy. The client’s content, which ranked well in organic results, was often too discursive for AI synthesis, covering topics broadly without the specificity that gets extracted for overview citations.

The Changes We Made

The GEO optimization program involved three main workstreams.

Content restructuring was the most significant effort. We went through the client’s highest-priority pages and revised them to surface key information more directly. This meant leading with clear answers before qualifications, restructuring headers to explicitly signal what each section answers, adding specific data points where the content had been making general claims, and cleaning up the information hierarchy so the most important content appeared higher.

Schema and structured data work added explicit machine-readable signals about content structure and entity relationships. We extended their existing schema implementation to include FAQ schema, HowTo schema where relevant, and more comprehensive Article schema with explicit author and publication metadata.

Entity authority building involved a content campaign to increase consistent mentions of the client’s brand in relevant topic contexts across credible external sources, strengthening the entity recognition signals that affect AI citation selection.

What Happened to Traffic

The results were not linear and they were not immediate. The first two months showed minimal measurable change. By month three, we started seeing the client’s content appearing in AI Overview citations for queries where it hadn’t previously. By month five, the pattern was clear enough to draw some conclusions.

AI Overview citation frequency increased meaningfully across the target query set. This didn’t translate immediately into proportional click traffic, which is the honest part of the story. For some queries, the AI Overview satisfied user intent without a click. For others, the citation created a click-through rate improvement because the overview established credibility that motivated deeper engagement.

Net organic traffic across the measured period was up, modestly but consistently. More meaningfully, the quality of traffic improved. Bounce rates dropped. Session durations increased. Conversion rates from organic traffic improved.

Generative engine optimization services done properly aren’t just about appearing in AI citations. They’re about improving content quality in ways that serve both AI systems and human readers, which produces benefits across multiple traffic channels simultaneously.

The Honest Caveats

Six months is a short observation window for SEO strategy evaluation, and we’re clear about that with the client. The full compound effect of improved content quality on traditional rankings is still developing.

Attribution is also imperfect. Some of the traffic improvements could reflect algorithm changes or competitive shifts rather than our GEO optimization specifically. We’ve tried to control for this in our analysis but acknowledge the limitation.

What we can say with reasonable confidence is that the direction of results is positive and the correlation with our specific interventions is strong enough to be meaningful. The client has continued the program past the initial six months.

Geo services as a discipline is still developing, and the practitioners doing this work honestly acknowledge that best practices are evolving. What we learned from this engagement is being built into our methodology, and the next client campaign will benefit from those learnings.

What This Means for Your Strategy

The practical takeaway for businesses evaluating GEO optimization is that the risk of engaging early is low and the potential benefit is real.

The content improvements required for GEO optimization are not neutral. They genuinely improve content quality in ways that benefit traditional rankings, user experience, and conversion rates, regardless of whether they produce AI Overview citations. The worst case is that you end up with better content that performs better across the board. The best case is that you capture AI citation visibility before your competitors figure out this is worth doing.

That asymmetry, limited downside, meaningful upside, is a reasonable argument for starting now rather than waiting for the discipline to fully mature.

`