How to show up in AI search (my field playbook after Marvin Chow’s chat)

Google’s Marvin Chow (VP, Consumer & AI Marketing) sat down with The Marketing Millennials to talk about showing up in AI search — worth your time because it reframes discoverability beyond classic rankings and into how AI systems select sources. 🎯 Watch the conversation.

I’m not recapping the chat. I’m giving you the playbook I use with clients to earn visibility across AI Overviews and chat assistants — and why each step matters. No tricks, just levers you actually control.

If you only have two minutes, start here.

What “showing up” actually means across AI surfaces

“Showing up” isn’t one thing. It’s a few:

• Inclusion or citation inside Google’s AI Overviews
• Being linked/quoted by assistants (ChatGPT, Claude, Perplexity, Bing Copilot)
• Appearing as a named source in summaries or sidebars

The common thread: systems choose sources they can parse, trust, and summarize without risk. You’re optimizing for selection and citation, not just position. 🧭

Signals you can control this quarter

Entity hygiene and schema

Your brand, people, and products need to exist as resolvable entities. That means consistent names, URLs, and identifiers across your site and key profiles — and explicit structured data.

• Ship Organization, Person, and Product (or Service) markup with strong sameAs and ids.
• Tie content to entities (author, organization, location) so it’s attributable and quotable.
• Audit schema gaps and conflicts; machines hate ambiguity. Start with the low‑noise, high‑impact fixes via the Schema Opportunity Analyzer. 🔎

If you’re new to this work, my deeper cut on aligning content with machine needs is here: Relevance engineering in AI search: my playbook after Mike King’s chat with Sean.

Crawlability, recency, and speed

Selection assumes discovery. Fix the pipes before you chase “AI tactics.”

• Resolve indexation blockers, prune thin/duplicative URLs, stabilize canonicalization.
• Refresh pages that should be timely; add change signals (lastmod, feeds, sitemaps).
• Improve performance and CLS so render-dependent content is reliably parsed. 🧱

If that sounds like a lot, scope it: a focused SEO Audit uncovers crawl-to-citation blockers in days, not months.

Citations andsource eligibility

Models won’t gamble on sources they can’t defend. Raise your selection odds with:

• First‑party data (surveys, usage telemetry, experiments) and original visuals.
• Clear authorship, bios, and editorial standards; connect experts to topics they actually work in.
• Targeted third‑party mentions on relevant, trustworthy sites. Avoid volume-for-volume’s-sake. One strong domain beats ten weak ones.

Evaluation and feedback loops

You can’t improve what you don’t measure. Track:

• How often you’re cited or linked inside major assistants for your target queries
• Which pages earn those mentions — and why
• Prompt variants that include/omit your brand to test robustness

Instrument this with the AI Visibility Checker and log changes weekly. When visibility shifts, validate whether the cause was content updates, link wins, or technical changes.

What I’d do this week

• Map your top 25 intents to pages, owners, and target entities. Kill or merge cannibal pages.
• Patch Organization/Person schema, align author pages, and sync sameAs across major profiles.
• Publish one definitive explainer with first‑party data or a compact original study; cite your method.
• Secure 3-5 relevant third‑party mentions that reference your unique contribution (not just your homepage).
• Set up weekly AI surface checks and annotate every site change so you can tie cause to effect.

Conclusion

AI search is a selection game. You earn selection by making yourself machine‑legible, low‑risk to cite, and measurably useful. Do the boring work well — entities, schema, crawl, and original contributions — and your visibility compounds. 📈

This isn’t theory. It’s the overlap of SEO fundamentals (discovery, structure), GEO realities (assistant selection), and AI practice (evaluation across models). If you want help operationalizing it, start with a quick read of my relevance engineering piece and run your domain through the AI Visibility Checker.