Harpal Singh (AI SEO / GEO)

ChatGPT thinks I’m a Top 20 AI SEO / GEO consultant // 70% of AI Search visibility is manufactured off-site. We unlock it. Founder @blimpp

[Harpal Singh (AI SEO / GEO) @CustomIntent]

We analysed 172 buyer prompts and 3,021 citations in women's intimate health supplements, then isolated the brand pages ChatGPT and Google AI Overviews actually cited.

There isn't one "perfectly optimised" page for AI search. The page AI wants changes with intent:

- Discovery prompts: 77% of brand citations went to editorial content.

- Shortlisting prompts: 48% editorial, 22% category, 24% product.

- Need-state prompts: 39% product pages

- Comparison prompts: evidence-led comparison content

- Purchase prompts: 45% of brand citations went directly to product pages.

We then analysed the cited PDPs themselves.

ChatGPT leaned toward hard product facts, suitability, commercial terms, verifiable references.

Google AIO leaned toward evidence explained, mechanisms & need-state questions.

But both need clinical validity, social proof and clear product data.

The result 👇

[Harpal Singh (AI SEO / GEO) @CustomIntent]

@batesbn PR really? Editorial is more important now, not less.

[Harpal Singh (AI SEO / GEO) @CustomIntent]

@JasonJh1319 If I had a big brand budget I’d be all over it, but can’t see the point for DR. An organic recommendation yes, a paid one no.

[Harpal Singh (AI SEO / GEO) @CustomIntent]

Top-of-funnel paid search no longer exists.

Discovery moved into ChatGPT need-state prompts & no amount of investment in AI Max will ever bring it back.

Demand Gen ads for truly cold audiences are also a weak substitute.

GOOG search income can only really go one way from here 📉

[Harpal Singh (AI SEO / GEO) @CustomIntent]

New default model in AI Mode since yesterday.

Google says Gemini 4 Hydrogen fans out across more subtopics & links a broader range of sites.

So any AI Mode citation tracking from before 12 Nov is measuring a different engine.

Remember to track this separately!

[Robby Stein @rmstein]

(1/6) The biggest model upgrade to Search yet has landed. Gemini 4 Hydrogen is here, and brings cutting-edge intelligence to all Search users. It’s now the new default model in AI Mode & being used for your hardest questions in AI Overviews.

> Breaks down complicated, multi-part Qs & precisely follows your prompts
> Reasoning advancements improve query fan-out to show more links to the web
> Much more visual, scannable & highly tailored responses (with comparison tables, charts & more)
> Rich understanding of all supported languages

I’ve been testing the model behind the scenes and it’s really incredible. It’s one of my favorite models I’ve ever used, and can’t wait for you all to try it! Here are a bunch of examples I found helpful 👇

[Harpal Singh (AI SEO / GEO) @CustomIntent]

@gaganghotra_ The problem with ‘enterprise’ AI SEO tools is that decision makers really don’t know any better and will gladly cough up $50k for the enterprise label, thinking that it’s what ‘serious’ brands do. Fear is driving these decisions.

[Harpal Singh (AI SEO / GEO) @CustomIntent]

@JohnFreelancing Measuring over a week is perhaps just noise. Try a 4 week trailing avg, running the same prompts multiple times across each engine.

[Harpal Singh (AI SEO / GEO) @CustomIntent]

@akhilpedia Bing feeds ChatGPT rather than Google AIO. That’s probably why.

[Harpal Singh (AI SEO / GEO) @CustomIntent]

@jasondavisseo And it’s really simple how to get into the AI pack. Proximity & reviews that fit the query context. That’s literally it.

[Harpal Singh (AI SEO / GEO) @CustomIntent]

@ThamiBenjelloun Need to crank those off-site listicles.

[Harpal Singh (AI SEO / GEO) @CustomIntent]

@JasonJh1319 Did the same on a recent trip to Rome. I’d say 90% of food & hotel recommendations were on point!

[Harpal Singh (AI SEO / GEO) @CustomIntent]

What’s going on with local SEO rn?

According to Reddit, one Phoenix law firm's average Maps grid rank went from 2.5 to 16.9 in late October.

It's still #1 in its market and everyone else dropped too.

One business owner on another Local Search subreddit found that 27 businesses shared the top-3 spots on his grid in September, and then 123 by end of October

Google hasn't confirmed anything but these fluctuations are becoming more common.

Before you panic:

1. Compare against competitors, rather than just last month.

2. Spot-check a few grid points on a signed-in phone.

3. Don't rip up your GBP until you’ve gone through 1 & 2.

[Harpal Singh (AI SEO / GEO) @CustomIntent]

@kenyondigital AI trackers aren’t dead, but paying $5k for essentially $100 worth of infrastructure will no longer be the norm

[Harpal Singh (AI SEO / GEO) @CustomIntent]

@Phy041 Industry dynamics. Anything health-related leans on authority sources more so than non-YMYL categories.

[Harpal Singh (AI SEO / GEO) @CustomIntent]

@ekuzevska Yes and that’s why so much money is going into prompt tracking and sentiment measurement rn

[Harpal Singh (AI SEO / GEO) @CustomIntent]

ChatGPT now does roughly 20% of Google's search volume but share of website traffic in October was 0.36% vs. 24.74% for Google (Ahrefs, 118k sites)

So that’s either a whole load of bot activity or it’s the final nail in the coffin for click-based attribution.

My guess is the latter.

[Harpal Singh (AI SEO / GEO) @CustomIntent]

Bit of both. People get the answer in ChatGPT based on its corpus, then convert down the line via Google or direct, so it never shows up as source: ChatGPT. They’re trying to fix this with in-agent checkout (real but still tiny). When that scales, there'll be no click to trace at all.

[Harpal Singh (AI SEO / GEO) @CustomIntent]

@tonyrestell “In AEO, backlinks barely matter”

Acquiring links on cited content actually is one of the core things you can do to improve your own AI visibility

[Harpal Singh (AI SEO / GEO) @CustomIntent]

@timgrassin 3 & 4 were probably what did it over 1 & 2.

[Harpal Singh (AI SEO / GEO) @CustomIntent]

@OpenAI The good bit is discovery topics. The rest is a bit meh.