Non-Commodity Content

This skill governs the first stage of content creation: getting real material out of the user's head and onto the page. Word count, formatting conventions, and page-type structure are handled elsewhere. The job here is substance.

Why this matters

Google does not judge content by whether a human or an AI wrote it. It judges the purpose and value of the finished page. AI Overviews and AI Mode are built on the same index, ranking systems, and quality systems as regular Search, so there is no separate technique that makes content perform in AI results. What works is what has always worked: useful, accurate content with clear structure, original thinking, first-hand experience, specific examples, clear authorship, and trustworthy sources.

Using AI to produce more pages faster is exactly the failure mode Google is filtering for. The value has to come from the user.

The core test

Before drafting and again after, ask: could any competitor in this category have published this page?

If yes, it is commodity content and it fails. The draft must contain something only this user could supply — a point of view, a lesson from client work, an original comparison, evidence from a real test, a specific example, or an answer that goes beyond what is already freely available.

Commodity: "7 Tips for Improving Your Sales Pipeline." Non-commodity: "The Five Pipeline Problems We Found While Auditing 60 HubSpot Portals."

One repeats common knowledge. The other adds experience, evidence, and a position.

Step 1: Interview before drafting

Never produce a draft on the first turn. The whole point of this skill is that the experience, judgment, and original thinking come from the user — if the model supplies them, the page becomes commodity content by definition.

Ask questions, wait for answers, then write. Ask in small batches rather than dumping twenty questions at once, and follow up when an answer is vague. "We help clients scale" is not usable material; "we moved three clients from 40 to 12 pipeline stages last year and close rates went up in two of them" is.

Always ask:

1. Who is this page for, and what decision are they trying to make when they land on it?
2. What do you know about this topic that most people writing about it don't?
3. What specific numbers, dates, client situations, test results, or examples can you cite?
4. What's your actual opinion here — including anything you believe that others in your field would push back on?
5. Who is the named author, and what makes them credible on this subject?
6. What sources will we cite, and are there any you specifically trust or distrust?

If it's a pillar page, also ask:

7. What's deliberately in scope, and what's out?
8. Which subtopics do you have direct experience with, and which are you only summarizing from general knowledge?
9. What existing pages should this link to, and what should link back to it?
10. What's currently ranking for this topic, and what does this page do that those don't?

If it's a landing page, also ask:

11. What exactly is being offered, and what happens after someone converts?
12. What are the top objections or hesitations, and how do you answer each one honestly?
13. What proof can you show — results, named customers, guarantees, screenshots, third-party validation?
14. What's the single action you want the visitor to take?

If it's a blog post, also ask:

15. What prompted this post — a client question, a pattern you noticed, an argument you want to make?
16. Is this responding to something else that exists, and if so, where do you agree or disagree with it?

Use judgment about order and phrasing. If the user has already answered something in conversation, don't ask it again — reflect it back for confirmation instead.

Step 2: Fill gaps visibly, never silently

Where the user hasn't supplied something, write a placeholder inline in the draft:

[NEEDS INPUT: specific number of portals audited]

Do not substitute a plausible-sounding statistic, case study, customer quote, date, or result. An invented specific is far worse than a visible blank — it reads as authoritative, survives editing, and destroys trust when a reader checks it. A blank gets filled.

The same applies to sources: don't cite a study, report, or article unless the user provided it or it has been genuinely verified.

Step 3: Flag what's still commodity

At the end of every draft, list:

• Every [NEEDS INPUT] placeholder.
• Every section that currently reads as generic and would fail the core test without the user's experience added to it.

Be direct about this. A section that summarizes common knowledge competently is still the thing Google is filtering out, and the user needs to know which parts of their draft are carrying no weight.

What not to spend effort on

These come up often and are wasted motion:

• Special schema or markup aimed specifically at AI search results.
• Chopping content into tiny "AI-friendly" chunks.
• Creating a separate page for every variation of a search query.
• Any tactic framed as an AI-search hack rather than as making the page more useful.

If the user asks for one of these, say plainly that AI features run on the existing ranking systems and the effort is better spent on substance.

Division of labor

Use AI for research, organizing ideas, drafting around supplied material, editing, repurposing, and speed.

Do not use it to supply experience, judgment, or point of view. Those come from Step 1, or they don't go in the page.