🔥 SEOs Diners Club #221: After Helpful Content: The Non-Commodity Content Era, Google Says "AI Search Is Still SEO," Microsoft Clarity Citations Goes GA

Hey everyone!

Last week we covered the five major updates Google shipped to AI Overviews and AI Mode: the "Expert Advice" feature surfacing Reddit quotes inside answers, the inline links, the hover previews, the subscription highlights, and the follow-up question suggestions. This week I went straight to the foundation underneath all those updates: the content discipline that comes after Helpful Content — what Danny Sullivan calls non-commodity content. I moved this story to the headline because for the rest of 2026 it's going to be the core of our client conversations.

Then Google made another move: it officially published a "what to do, what to skip" guide for generative AI search. In the same week, Microsoft made the AI Citations dashboard inside Clarity (its free analytics tool) generally available. You can now see, for free, how many times your pages get cited in AI responses and what share of authority you hold against your competitors on the same queries.

New data was published on how ChatGPT's crawler behaves, Anthropic shipped Claude for Small Business, and Google Search Console had a brief data logging bug in the Discover report. Next week brings Google I/O 2026 (May 19–20) and Google Marketing Live 2026 (May 20) back to back.

The undertone of this week's issue is clear: we're working for the moat, not for the volume. It's the moment to drop commodity content and move to non-commodity. Google may say "AI search is still SEO," but ChatGPT, Perplexity, Claude, and the Gemini App are playing a different game. The visibility work of 2026 requires a separate playbook for each engine.

Grab your coffee. Let's dive in.

🎯 After Google Helpful Content — The Non-Commodity Content Era

Google Search Liaison Danny Sullivan's December 2025 Search Off the Record episode split the industry in two: on one side commodity content, on the other non-commodity content. This split is the direct successor to the Helpful Content system launched in 2024. In 2026 it has become the shared yardstick of both Google rankings and AI citations. Think of it the way a commodities market works: standardized goods like wheat or gold are interchangeable, but a one-off Picasso painting cannot be substituted. Non-commodity content is the one-off Picasso of the open web.

The One-Sentence Test

When you hand ChatGPT a title and ask for 800 words, does the output read close to yours? If yes, you've produced commodity content. If no, it's non-commodity. Anything an AI can write is now commodity by default; the only thing that's still non-commodity is what no one else could write.

Sullivan's own example nails it. For a real estate agent, "7 Tips for First-Time Homebuyers" is classic commodity content (pre-approval, location, budget). The counter-example: "We Waived the Inspection and Saved $15,000: What We Found in the Sewer Line." That second one is non-commodity. The agent personally crawled the sewer line, saw it was PVC and not concrete, and made the call in the middle of a bidding war. ChatGPT can't write that — because no one had written it before.

The Six-Dimension Scoring Framework

At Stradiji we score non-commodity content on six dimensions: Proprietary Evidence (25%), Firsthand Experience (20%), Specificity (15%), Point of View (15%), LLM Moat (15%), and Information Gain (10%). Total is out of 100. Above 70 is non-commodity, 40–69 is mixed, below 40 triggers a commodity warning.

Three Pressures, All at Once

Three pressures explain why this matters now. First, generative AI is producing millions of commodity pieces every day. Where there used to be 50 articles on a topic, there are now 5,000. The marginal value of commodity content is approaching zero. Second, Helpful Content has been directly rewarding "firsthand experience" signals since 2024. Third, AI engines are dependent on third-party authority signals — which is exactly why commodity content can't make it onto their citation lists.

Mert's Note

We've been doing SEO at Stradiji since 2009, and 2026 is the most aggressive year I've seen for the shift from "publish a lot" to "publish less but make it non-commodity." The first thing we tell clients now is this: start asking your team "what cases did we work this week, what numbers did we collect, what contentious calls did we make?" Don't produce content before those answers come in. Once they do, the content tilts toward the non-commodity side on its own.

The client conversation to open this week is: "What percentage of our content portfolio is on the non-commodity side? What percentage is throwing commodity warnings?" If the answer is unclear, start the six-dimension scoring exercise this month. If the answer comes back "80% commodity," don't panic — that's where most sites are starting. Producing one non-commodity piece per month is more efficient than three commodity pieces per week. Your editorial calendar metric should not be "posts published," but "non-commodity pieces published."

I've written a long-form article alongside the newsletter on this. It includes a step-by-step playbook for moving from commodity to non-commodity, a case from a Stradiji marketplace client (three pages, four weeks later, average Google position dropped from 14.3 to 6.1), and a validator scoring tool. The full link is at the end of the issue.

This Week's Standout Resource: HubSpot Marketing Team's Perplexity Study Page

While digging into this topic I came across a genuinely useful resource: HubSpot's marketing team published a dedicated Perplexity study page as part of their "How to Rank on Google" email campaign. The page walks through the commodity vs. non-commodity distinction with live examples and embeds a grader right inside. You paste your own content and get an instant score on a six-dimension framework. We at Stradiji tested it this week on our last 10 posts at Stradiji.com and on two client sites — the results lined up with our own six-dimension framework. So this isn't just theory; it's a measurable method.

Practical recommendation: after reading the long-form article alongside the newsletter, open HubSpot's Perplexity page and paste your three most important pages into the grader. If the score is below 40, put that page on a rewrite plan. HubSpot's companion tool, the AEO Grader, measures AI visibility at the brand level. Using both gives you a clean picture at the page level and the brand level.

🔄 Google's New Generative AI Guide — "AI Search Is Still SEO"

Google published a new guide on Search Central on May 14, 2026, titled "Optimizing your website for generative AI features on Google Search." The guide is an expanded version of the 2025 documentation and gives Google's official answer on two acronyms the industry has been debating: AEO (answer engine optimization) and GEO (generative engine optimization). Google's response is unambiguous: "From Google Search's perspective, optimizing for generative AI search is optimizing for the search experience, and thus still SEO." (Search Engine Journal, ALM Corp)

The most notable part of the guide is the list of practices Google tells site owners they can skip:

• No llms.txt file needed. Google says explicitly that you don't need to create machine-readable files, AI text files, or any special markup or Markdown to surface in generative AI search.

• No content chunking needed. "Our systems are able to understand the nuance of multiple topics on a page and show the relevant piece to users." Translation: don't break your page into 100-word AI-friendly boxes.

• No special schema needed. No new schema types specific to AI features are documented by Google.

At the same time, the guide reaffirms the value of classic SEO: AI features are "rooted in our core Search ranking and quality systems," and content is pulled from the Search index.

Important Caveat: This Is Only About Google — Other AI Engines May Still Use These Practices

I want to underline this clearly: Google's guide only applies to Google Search's AI Overviews and AI Mode. ChatGPT, Perplexity, Claude, and the Gemini App (the standalone consumer app that runs separately from Google Search) are built on different infrastructure. A few concrete examples:

• llms.txt: Google considers it irrelevant; however, Anthropic's official documentation has formally acknowledged llms.txt. Perplexity is also reading llms.txt experimentally. "Google doesn't want it" doesn't translate to "no AI wants it."

• Schema (structured data): Ahrefs' 1,885-page study from last month showed that schema alone doesn't meaningfully lift AI citations. The sample is limited and the conditions matter. But for traditional rich results and "People Also Ask," classic schema remains essential.

• Content chunking: The way ChatGPT's crawler reads a page, content split into short, clean blocks is easier to cite. Q&A blocks, brief definition boxes, and lists are favored in ChatGPT answers.

Mert's Note

This guide crystallized two distinct strategy layers for us at Stradiji.

Layer 1 — for Google Search's AI Overviews and AI Mode: keep your classic SEO work going. Helpful Content guidelines, E-E-A-T, title and meta description optimization, internal linking, Core Web Vitals. Google's guide is explicit: "You don't need to do anything new; just keep getting better at the fundamentals."

Layer 2 — for ChatGPT, Perplexity, Claude, the Gemini App: crawler-friendly architecture (server-side rendering), chunkable block layouts (semantic triples, definition boxes, lists), experimental llms.txt setup, and third-party mentions (Reddit, Quora, industry publications) still matter. None of this is in Google's guide, but it moves the needle on other engines.

Here's the clean message to give clients this week: "Half of your SEO budget goes to Google; the other half goes to multi-engine." The era of one strategy winning everywhere is gone. Google's own guide just made the split official.

Use the GEO Scorecard to measure your position engine by engine.

🔄 Microsoft Clarity AI Citations Dashboard Now Free for Everyone

On May 13, 2026, Microsoft made the AI Citations dashboard inside Clarity — its free analytics tool — generally available. Originally launched as a beta in February, the dashboard rolled out to all Clarity users after three months of refinement. (Microsoft Clarity Blog, PPC Land)

The dashboard reports two core metrics:

• Page Citations: the number of times pages from your domain are referenced in AI-generated answers during the selected period. Multiple citations within the same answer are counted separately.

• Share of Authority: your domain's share of total citations among other cited domains on the same set of queries. A competitive view: if ChatGPT issued 100 citations on a topic, how many are yours?

Clarity is already free, and there's no extra fee for the new dashboard. All you need to do is install the Clarity tag and wait for the data to accumulate.

Mert's Note

We've been telling clients for months: "Measure your brand visibility in AI answers." But for most clients in the Turkish market, the budget for GeoGenie or a comparable paid measurement platform wasn't there. Microsoft Clarity's new dashboard closes that gap, for free.

Three steps this week: (1) Check Clarity installation on client sites; install it this week where it's missing. (2) Open the AI Citations dashboard and document the 28-day baseline. (3) If Share of Authority is low, list the queries where you trail competitors. Start content work next month to close those gaps.

One important caveat: Clarity's data focuses on Microsoft's AI ecosystem (primarily Bing Copilot, Microsoft 365 Copilot). ChatGPT, Perplexity, and Google AI Overviews need to be measured through different sources. Even so, as a free starting point, the AI Citations dashboard is the most practical addition of 2026.

🔄 ChatGPT's Crawler Sends 3.6x More Requests Than Googlebot

Search Engine Land published a deep-dive on May 14, 2026, explaining the underlying architecture of ChatGPT Search. A separate study from the same weeks showed that the ChatGPT-User crawler revisits high-traffic pages roughly every 2.4 days. An analysis of more than 24 million proxy requests across over 78,000 pages found that ChatGPT-User makes 3.6 times more requests than Googlebot. (Search Engine Journal)

Two Distinct OpenAI Crawlers

OpenAI runs two separate crawlers: ChatGPT-User, the retrieval crawler that pulls pages in real time when a user asks something — this is the bot that determines whether you show up in ChatGPT answers. GPTBot, the training crawler that gathers data for model training. The two have different purposes, and you need to manage them separately.

Technical Detail: Server-Side Rendering Is Mandatory

Unlike Googlebot, GPTBot doesn't use a full browser and doesn't render JavaScript. It sees your raw HTML response. Server-side rendering (SSR) is no longer optional, it's mandatory. Content loaded only by client-side JavaScript won't appear in ChatGPT answers.

Mert's Note

This data has three practical consequences. First, analyze your server logs regularly. Knowing which pages ChatGPT-User pulls — and how often — is now a direct measure of SEO performance. Second, test your SSR setup. Sites running headless CMS or JS frameworks (Next.js, Nuxt, SvelteKit) need to answer the question "what does the crawler see in the raw HTML?" Third, manage robots.txt rules separately for GPTBot, ChatGPT-User, OAI-SearchBot, ClaudeBot, and PerplexityBot. Blocking all of them kills visibility; opening all of them spikes server load. Run the traffic analysis and decide based on data.

Produce a one-pager for clients this week: "Our AI crawler access policy." Document your decision for each crawler (allow / block / limited) and the reason. Make it a living document you'll revisit in three months.

🔄 Three Big Moves from Anthropic: Claude for Small Business, Microsoft 365, AWS Platform

On May 13, 2026, Anthropic introduced the Claude for Small Business package. The bundle ships ready-to-run workflows for Intuit QuickBooks, PayPal, HubSpot, Canva, Docusign, Google Workspace, and Microsoft 365. One-click installation, aimed at accelerating small-business AI adoption. (SiliconANGLE)

The same week, Anthropic shipped Microsoft 365 add-ins for Excel, PowerPoint, and Word into general availability and moved the Outlook add-in to open beta for paid plans. Third move: Claude Platform on AWS, giving access to the Claude API through Anthropic-managed infrastructure with AWS billing and IAM authentication.

Mert's Note

Read together, Anthropic's strategy comes into focus: horizontal expansion in the enterprise market. Claude for Small Business competes with ChatGPT's "one tool, many jobs" positioning; the M365 add-ins compete with Microsoft Copilot on the same field; the AWS platform offers cloud-provider independence on the distribution side. The SEO and GEO consequence: your clients' AI strategy is no longer tied to a single model. A client's marketing team might be on ChatGPT, the legal team on Claude, and the sales team on Microsoft Copilot. Build your visibility work as multi-engine from the start.

🔄 Google Search Console Discover Report Bug on May 7–8

Google confirmed that between May 7 and May 8, 2026, a data logging bug affected the Discover performance report in Search Console. The bug caused fewer clicks and impressions to show up in the report, but it didn't affect actual Discover rankings. Google described it as "a data logging issue only." (Optimixed)

Mert's Note

Two takeaways. First, don't flag instant swings as "performance drops" to clients. A weekly comparison drop might be real; a one-day dip usually isn't. Second, Search Console's data issues keep getting updated. The long-running impression logging issue between May 13, 2024 and April 27, 2026 was also recently closed. Start adding a standard footnote in monthly reports: "Search Console data was corrected after this date; earlier comparisons are no longer valid."

🔄 Next Week: Google I/O 2026 and Google Marketing Live 2026

Two major events back to back. Google I/O 2026 runs May 19–20 in Mountain View. Expected headline: Gemini 4. A 10-million-token context window, native multimodal capabilities (no transcription preprocessing), Android 17 with a unified "Adaptive Everywhere" approach, and an early look at Android XR glasses. (Android Authority)

Google Marketing Live 2026 follows on May 20 as a virtual broadcast. AI-powered campaign tools, agentic commerce capabilities, and a new performance era on YouTube are the headline tracks. (ALM Corp) Based on the 2025 pattern, Gemini updates announced at I/O reached the ad tools within hours; the same pattern is expected this year.

Mert's Note

Next week's issue will break down every important announcement from both events one by one. As prep, draft a note answering three questions: (1) Does Gemini play a direct role in my clients' ad strategy? (2) If agentic commerce gets announced, which of my clients' product catalogs are ready? (3) If Android XR glasses get introduced, how should my local-search-focused clients reposition? Answer these in advance so you can publish your commentary within hours of the event.

🤖 From the AI World

Perplexity Comet Arrives on iOS: Perplexity's Comet browser opened to all iOS users this week. Available on all four major platforms now: Mac, Windows, Android, and iOS. The Personal Computer assistant ties together local files, apps, and browser context.

Anthropic's "Evil" Explanation: Anthropic explained that earlier blackmail-attempt behaviors from Claude were rooted in "evil AI" depictions in training data. Debate is open on how fictional patterns in training data leak into model behavior.

ChatGPT Memory Sources: Memory Sources, shipped alongside last week's GPT-5.5 Instant release, is now live on the web for Plus and Pro users. Past conversations, files, and Gmail content are processed by the search tool to produce more personalized answers.

Google Spam Policy Reminder: The back button hijacking penalty starts June 15. We've covered this for three weeks running — this is the last good window to run your tests.

Competitive Analysis Inside ChatGPT Ads: Adthena introduced its platform offering competitor and performance visibility inside the ChatGPT ads ecosystem. The emergence of third-party tooling in AI search advertising is a sign the category is maturing.

✅ What To Do This Week

1. Read Google's new generative AI guide with your full team. Draft two separate strategy notes: "for Google" and "for other engines."

2. Check Microsoft Clarity installation across all client sites. Install it this week where it's missing. Start collecting the 28-day baseline on the Citations dashboard.

3. Analyze your server logs for ChatGPT-User, GPTBot, ClaudeBot, PerplexityBot, and GoogleOther crawler behavior. Which pages get pulled how often, and which ones don't get pulled at all?

4. Test your server-side rendering (SSR) setup. Does JavaScript-loaded content appear in the raw HTML? If not, your page is invisible to GPTBot.

5. Add a footnote about the May 7–8 Search Console Discover bug to client reports. Prevent misreadings in month-over-month comparisons.

6. For US and UK clients, evaluate the ChatGPT Ads Manager beta sign-up opportunity.

7. Prepare a follow-up note for Google I/O 2026 (May 19–20) and Google Marketing Live 2026 (May 20). Three prep questions: How does Gemini 4 affect my ad strategy? If agentic commerce gets announced, which of my catalogs are ready? How does Android XR change local search?

8. Produce a one-pager titled "Our AI crawler access policy." Document the decision (allow / block / limited) and the rationale for each crawler.

9. Complete back-button-hijacking testing this week. Four weeks left to the June 15 enforcement; third-party ad scripts carry the highest risk.

10. Use the non-commodity content framework from this week's headline to score your clients' last 30 pages. Build a rewrite calendar for the pages throwing commodity warnings. For the full framework and step-by-step migration playbook, read the long-form article alongside the newsletter.

👋 Closing

This week the headline was non-commodity content as a discipline. Alongside that, Google answered the generative AI search question with "AEO and GEO are still SEO," and Microsoft Clarity made its AI Citations dashboard generally available. New data was published on ChatGPT's crawler behavior, Anthropic shipped Claude for Small Business, and Search Console had a brief Discover report bug.

The overarching message is clear: in 2026, we're working for the moat, not for the volume. As commodity content production gets cheaper, it loses value. Non-commodity content, on the other hand, earns citations across every engine. At the same time, building one strategy for one engine isn't enough. "Keep doing the SEO fundamentals" is the right message for Google, but ChatGPT, Perplexity, Claude, and the Gemini App are playing a different game. Split your SEO budget into two layers: one for Google, one for the other engines. Then lay non-commodity content discipline underneath both as the third layer.

Next week we have Google I/O 2026 and Google Marketing Live 2026. We'll go through Gemini 4 and agentic commerce announcements together.

See you next week — take care of yourselves!

Mert Erkal Stradiji | SEO, GEO & Conversion Optimization

Support the newsletter: if you find this content useful, buy me a coffee. ☕

About Mert Erkal Founder of Stradiji. 15+ years in SEO and GEO consulting for corporate clients globally. Author of SEOs Diners Club (English) and Dijital Pazarlama Notları (Turkish).

• Newsletter: seosdinersclub.beehiiv.com

• Twitter: twitter.com/merterkal

• LinkedIn: linkedin.com/in/merterkal

• YouTube: youtube.com/@stradiji