Head of Growth, Canadian B2B SaaS firm (Series B, ~80 employees, vertical SaaS)

A Series B Canadian vertical-SaaS firm was watching competitors get cited consistently in ChatGPT, Perplexity, and Google AI Overviews answers for its core category queries — while the client appeared in roughly 8% of probes. We ran a 7-month AEO/GEO program focused on direct-answer formatting, fact density, byline schema, and original-research publication.

Service & engagement

The challenge

The client's category was being increasingly answered directly in AI engines rather than via traditional SERPs. Their internal team had measured citation share on a 50-query basket: client appeared in 8% of ChatGPT responses, 4% of Perplexity, 11% of Google AIO. The two largest competitors appeared in 60–80% of probes across the same engines. The marketing team had no playbook for AEO/GEO and was tempted to write it off as a fad.

Approach

1. 1
   Established a baseline citation-share measurement across 50 category queries × 4 engines (ChatGPT, Perplexity, Gemini, Google AI Overviews), tracked monthly throughout the engagement.
2. 2
   Rewrote the top 30 highest-traffic pages to lead with explicit direct-answer paragraphs (3–5 sentences answering the query subject without pronoun openings), followed by structured detail.
3. 3
   Added named-author bylines + Person JSON-LD across the content library, tied to a real bio page with credentials, publications, and external citations.
4. 4
   Increased fact density to 3+ verifiable facts per 500 words across the top 30 pages — currency amounts, percentages, named regulators, durations, dates — replacing softer claims.
5. 5
   Published two pieces of original research over months 3 and 5: hand-collected data from a 200-customer survey + an internal analytics study, both formatted with a methodology section, dataset summary, and Article + Dataset schema.
6. 6
   Re-allowed and verified all major LLM crawlers in robots.txt (ChatGPT-User, OAI-SearchBot, PerplexityBot, ClaudeBot, GoogleOther) and shipped a clean llms.txt + llms-full.txt manifest at month 2.

Outcomes

Reflection

AEO/GEO is no longer optional in categories where the AI engines are answering the query directly. The lever set is finite and specific: direct-answer formatting, fact density, byline schema, original research, and LLM crawler access. Programs that treat AEO as 'just write better content' don't move the citation-share needle; programs that systematically address the five levers move it within 90 days. The challenge is that classical-SEO ranking and AI-engine citation are now distinct disciplines — but the same content team can run both with a clear playbook.