Enterprise B2B SaaS AI Search Visibility Case Study

The client identity is withheld under confidentiality. The figures are published as bounded observations from the engagement rather than as universal performance promises.

The company entered the engagement with substantial brand awareness, a strong domain, a large content footprint, new leadership and a major website migration ahead. Traditional SEO signals suggested a credible category presence. The initial AI visibility research showed a different condition.

The starting finding did not prove that the brand was absent from every AI answer. It showed that the brand was absent from this defined sample of category, problem, comparison and solution questions. The working question became: which technical, content and source conditions were preventing the company's existing authority from translating into buyer-facing AI visibility?

The research modeled approximately 100 representative buyer profiles across a seven-stage decision journey. Instead of relying on broad prompts such as “What are the best revenue intelligence platforms?”, the study covered commercially specific decisions.

This produced a buyer-question corpus grounded in commercial decisions rather than search volume alone.

The initial benchmark recorded 171 unbranded answers and reviewed brand mentions, recommendations, context, cited domains, competitor positioning, repeated claims, content formats and buyer-stage differences. The client appeared in none of those answers.

The zero-mention finding was treated as a starting observation, not the diagnosis. The team used the answer and source records to investigate why competitors remained eligible and what evidence the client lacked at each stage.

The roadmap covered crawler access, indexation, canonicalization, internal architecture, crawl allocation, structured content, structured data and migration planning. AEO extended the retrieval environment; it did not replace technical SEO.

The research identified 48 commercially important questions without adequate corresponding content. They were selected because they affected real buying stages, not because a keyword tool reported high monthly volume.

The prioritization model connected keyword demand, buyer intent, AI prompts, competitor visibility and commercial importance. It gave low-volume but high-consequence questions appropriate weight.

The team mapped the sources shaping category answers across vendor sites, review platforms, publications, comparison sites, analyst-style content, communities, educational resources, product documentation and original research.

This changed the work from a content calendar into a source-and-evidence strategy. The question was not only what the client should publish, but where trustworthy evidence needed to exist for buyers and retrieval systems to encounter a consistent account of the brand.

Pricing prompts exposed a weakness that a top-line visibility score would have hidden. In the pricing-stage analysis, a major competitor received citations in approximately 30.6% of measured responses while the client received citations in approximately 4.6%.

The gap led to recommendations covering pricing context, purchase criteria, value justification, packaging, implementation expectations, total cost and independent pricing evidence.

The page and comparison counts describe the roadmap and priority set. They should not be read as a claim that every identified asset was published within one period.

The program did not end with a higher visibility observation. It revealed where the brand was still losing. Pricing-stage citation weakness became a new priority even while broader recommendation visibility was strong.

A dashboard reports movement. Strategy turns the movement into the next controlled decision.

1. Business intelligence: company, product, market, competitors, differentiation and commercial goals.
2. Audience and buyer intelligence: ICPs, roles, jobs, objections, risks, criteria and journey stages.
3. Search and prompt intelligence: keywords, SERPs, buyer questions, conversational prompts and sales questions.
4. AI visibility benchmarking: mentions, recommendations, citations, context, competitors, engines and stages.
5. Retrieval and source analysis: cited domains, repeated evidence, influential environments and competitor support.
6. Technical retrieval audit: bots, robots, rendering, responses, crawl allocation, architecture, schema and migrations.
7. Evidence-gap analysis: the facts buyers and systems need but the public environment does not adequately prove.
8. Intervention: technical fixes, content, comparisons, research, reviews, PR, communities and structured data.
9. Retesting: repeat the same commercially relevant question set.
10. Learning: turn findings into the next hypothesis, change and evidence requirement.

The client could be discovered through ChatGPT, Google AI Mode, AI Overviews, Gemini, Claude, Perplexity, conventional search, communities, video, reviews and publications. The objective was not to manipulate one model. It was to build an information environment in which the brand was understandable, retrievable, credible, differentiated and well evidenced wherever enterprise buyers researched the category.

Confidentiality: The client is blinded. Identifying details, proprietary prompts, source exports and internal analytics are not published. Outcome boundary: These results document one engagement and do not guarantee future visibility, citations, traffic or revenue.

What did the initial AEO benchmark find?

The blinded enterprise B2B SaaS client appeared in zero of 171 unbranded AI answers in the initial defined sample.

Did the client's AI visibility increase from zero to 82%?

That conclusion would overstate the evidence. The initial 171-answer benchmark and later 82% mention-rate scan used different prompt sets, so the results are reported as separate observations rather than a controlled pre-post change.

What did the 756-prompt Google AI Mode study find?

The client was recommended in approximately 45.4% of responses, compared with approximately 38% for its primary direct competitor within that specific study.

What organic-search results were observed?

Organic traffic increased 20.8% quarter over quarter, and the number of keywords ranking in positions 4 to 10 increased approximately 40% during the analyzed period.

What technical problems did the audit identify?

Some major AI crawlers received HTTP 403 responses, and one subdomain consumed approximately 83% of observed Google crawling activity.

Can these AEO results be guaranteed for another company?

No. The results document one blinded engagement with specific starting conditions, question sets, competitors, platforms and measurement periods.