GEO, AEO and SEO: a July 2026 retrospective

By Sumith Parambat DamodaranPublished in TechnologyOctober 08, 20264 min read

GEO and AEO are useful lenses for planning content, but neither removes the need for a useful, accessible website. The practical question is whether someone can find a trustworthy answer about your product and follow it back to evidence.

Retrospective theme date: 17 July 2026. First published: 8 October 2026. This is a newly written reflection organised as part of a July–October series; it was not published on the earlier theme date. Sources and product information were reviewed on 8 October 2026.

Three labels, one customer problem

I use SEO for discoverability in search, AEO for answering a question clearly, and GEO for becoming a useful source in generated responses. These are working distinctions, not universally agreed categories. A single page can serve all three goals.

Myriam Jessier’s Fundamentals of GEO workshop at BrightonSEO introduced the distinction between answering questions and becoming a source. It also separated what a model may recall from training from information retrieved while answering. That distinction is useful, with a caveat: an answer alone cannot establish where every statement came from. Asking a model to score a brand is an exploratory exercise, not an independent reputation audit.

For a small product team, I would turn this into a practical content review. Take one buying decision, such as choosing a booking system for a small studio. List the facts a buyer needs: supported booking rules, payment options, cancellation handling, integrations and limitations. Then check whether those facts are available on maintained pages, with consistent terminology and an obvious owner.

A worked example

Imagine a fictional booking product called StudioLedger. Its homepage says it is flexible, powerful and easy. Those adjectives give a buyer little basis for a decision.

A more useful page explains that recurring classes are supported, names the available payment integrations, describes cancellation deadlines and states what happens when a class is full. If the product does not support a capability, the page should say so. This example is illustrative; it describes no real vendor or customer.

The next step is evidence. A documented workflow or a reproducible demonstration is more useful than repeating the claim across ten pages. A comparison should define its criteria and review date. A case study should explain how an outcome was measured and have permission for publication.

Keep the brand facts coherent

The workshop’s discussion of brand identity prompted a simple editorial rule: maintain one authoritative home for facts that frequently change. Product names, availability, support policies and specifications should not contradict each other across a website.

Independent sources still matter. A brand cannot control every review or every generated response. It can correct factual errors in its own material, support credible claims and avoid creating conflicting versions of its story.

Google’s current guidance keeps SEO fundamentals relevant to its generative search features. It also says special AI files are unnecessary for visibility in Google Search. That is guidance about Google, not a guarantee about every assistant.

My first experiment

I would improve one decision page before commissioning a large content programme. Record the missing facts, publish the corrections, and check both whether people can complete the task and whether monitored answers represent those facts accurately. More mentions would be interesting; fewer misleading answers and better qualified enquiries would be more useful.

The objective is a better decision for the reader. AI visibility is a signal to investigate, not a reason to fill the website with repetitive answers.

Sources and attribution

Myriam Jessier, Fundamentals of GEO, BrightonSEO 2026 workshop, supplied slide deck, slides 5, 8–16. Concepts discussed here are paraphrased; the examples and proposed workflow are original.
Google: optimizing for generative AI features in Search, checked 8 October 2026.

Illustrative cover image: existing site photograph by Pankaj Patel, reused from the Flutter VSCode extensions article. It is not workshop or product imagery.