17:51Caleb Ulku · Tutorial
TRAIN Any AI To Recommend You (This Study Proves How)
How an Anthropic study on 250 documents becoming part of AI memory became a local-business content strategy.
June 12thThe argument in one line.
AI agents have compressed the entire customer funnel into one automated conversation, so websites must now convince a machine before they ever get a chance to convince a human.
Read if. Skip if.
READ IF YOU ARE…
- A business owner or marketer who has invested in SEO and wants to understand how AI search changes the game.
- Someone building or rebuilding a website who needs a concrete checklist for making it agent-readable.
- A founder who sells services and wants to understand how AI agents evaluate and shortlist vendors.
- Anyone trying to forecast where to put marketing effort in the next 12-24 months.
SKIP IF…
- You want a deep technical implementation guide -- this is a strategic overview with actionable direction, not step-by-step code.
- Your business has no web presence and relies entirely on referrals or offline channels.
The full version, fast.
AI agents are already doing the product research that humans used to do, and they evaluate websites on five mechanical signals: structured data, content clarity, API access, third-party brand mentions, and freshness. A site that fails any one of these signals gets eliminated from the shortlist before a human ever sees it. The businesses that address these signals now, while most competitors are still unaware, lock in a recommendation flywheel that compounds -- more agent referrals drive more traffic, more traffic builds more authority, more authority earns more referrals. The early-mover window is months, not years.
Free for members
Chat with this breakdown — free.
Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.
Create a free account →Where the time goes.
00:00 – 00:43
01 · Cold open: the shift stated
Hook premise delivered in full: AI agents are already buying, companies they skip will never know they were in the running.
00:43 – 02:05
02 · Chapter 1: The Internet Just Changed Again
Historical framing: three waves (desktop SEO, mobile, AI agents). Early movers in each locked in advantages that held for years.
02:05 – 03:34
03 · Chapter 2: The New Customer Journey
Old funnel vs. new funnel. Human delegates research to AI agent; full awareness-consideration-decision cycle compresses into one conversation. Human never visits a site.
03:34 – 06:37
04 · Chapter 3: How AI Agents Actually Shop
Five signals agents evaluate: structured data, content clarity, API compatibility, brand signals, freshness. Accessibility equals AI visibility.
06:37 – 09:18
05 · Chapter 4: The Agent-Ready Playbook
Six concrete steps: schema markup, rewrite for clarity, open APIs, build brand signals, keep content current, create category-ownership content.
09:18 – 11:08
06 · Chapter 5: The First-Mover Window
Gartner 2028 prediction, 12-18 month adoption timeline, recommendation flywheel mechanics, urgency close.
Lines worth screenshotting.
- AI agents now handle awareness, consideration, and decision inside a single conversation -- the human only picks from a pre-filtered shortlist.
- A website that cannot answer what you sell, who it is for, what it costs, and how it works in plain machine-readable language does not make the agent shortlist.
- Schema markup is the single highest-ROI change you can make right now -- it takes an afternoon and makes every AI agent on the internet able to read your offer.
- The same ARIA tags and clean HTML that make sites accessible to people with disabilities are now the exact signals that make sites visible to AI agents.
- AI agents cross-reference your entire web presence before recommending you -- if no one else on the internet is talking about you, the agent does not trust you.
- Mobile took five years to reach mainstream adoption; AI agents will do it in 12 to 18 months, which compresses the early-mover window from years to months.
- Stale content is a trust killer for agents -- last-updated dates and quarterly refreshes are now a competitive signal, not a housekeeping task.
- Specificity trains AI: the more precisely and consistently you describe your category, the stronger the association AI builds between your brand and that niche.
- Once AI agents learn which brand to recommend in a category, it becomes a self-reinforcing flywheel -- recommendations drive traffic, traffic builds authority, authority earns more recommendations.
- Most businesses will not act on this even after watching a video about it -- which is why acting now, imperfectly, beats waiting for the perfect strategy.
What AI agents actually look for when they evaluate your site
WHAT TO LEARN
AI agents eliminate candidates mechanically before a human ever sees the shortlist, and the five signals they use are concrete, fixable, and mostly ignored by your competitors right now.
- Structured data (schema markup / JSON-LD) is the highest-leverage single change you can make -- it takes an afternoon and tells every AI agent exactly what you offer, at what price, and for whom.
- Content clarity beats brand storytelling for agent audiences: every key page needs to answer four plain-language questions -- what you sell, who it is for, what it costs, and how it works.
- Accessibility infrastructure and AI readability are the same thing -- ARIA tags and clean semantic HTML serve both disabled users and the agents that evaluate your site.
- Third-party brand presence is a trust signal, not a vanity metric -- agents cross-reference mentions, reviews, and citations before recommending you, so a site with no external footprint gets filtered out.
- Freshness is evaluated mechanically -- adding last-updated dates and refreshing core content quarterly shifts how agents assess your site's reliability.
- Specificity in how you describe your category trains AI systems to associate your brand with that niche, and early associations compound through the recommendation flywheel in a way that makes late movers fight uphill.
Terms worth knowing.
- Schema markup / JSON-LD
- Structured data code added to a webpage that tells search engines and AI agents exactly what the page is about, what is sold, at what price, and for whom. Without it, AI agents must guess, and they often skip you instead.
- AI agent
- An AI system that acts autonomously on a user's behalf -- searching the web, comparing options, and returning a shortlist or completing a transaction without the human visiting any site directly.
- Universal Commerce Protocol
- A Google initiative that lets AI agents handle the full buying journey from discovery through checkout, rewarding sites with clean APIs that agents can plug into directly.
- ARIA tags
- Accessible Rich Internet Applications attributes embedded in HTML that describe page elements to screen readers for people with disabilities -- now also read by AI agents as structural signals.
- First-mover window
- The early period after a major platform shift when adopters can lock in compounding advantages before the market normalizes and those advantages become table stakes.
Lines you could clip.
04:21
“The entire funnel -- awareness, consideration, decision -- got compressed into one AI conversation.”
04:35
“The website didn't need to convince a human. It needed to convince an agent.”
04:34
“Accessibility for humans became visibility for AI.”
10:24
“Most of your competitors won't watch this video. Most won't even take action even if they do. That's your advantage. Use it.”
Word for word.
Read-along
Don't just watch it. Burn it in.
See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.
metaphoranalogy
In the next twelve months, millions of people will stop searching for products and start letting AI agents buy for them. The companies they choose will see a massive surge in revenue. The companies they skip will never even know they were in the running.
This is the biggest shift in how customers find businesses since the invention of Google, and almost no one is talking about it. I'm Neil Patel.
I'm the cofounder of NP Digital. We're the performance marketing agency of the year by Ad Age. And in this video, I'm gonna show you how AI agents actually shop, how they scan, compare, and decide who gets the sale.
Then I'll walk through exactly what to change on your website so these agents choose you every time.
There have been exactly two massive shifts in how people find businesses online. The first was the desktop Internet. Google launch and suddenly, if your business showed up on page one, you won.
The companies that learned SEO early dominated for a decade before the competitors even understood what happened. The second was mobile. Smartphones change everything.
The businesses that went mobile first, responsive sites, fast load times, apps, captured market share, they still hold today. The ones that waited, a lot of them never recovered.
Right now, we're standing at the beginning of the third shift. This one's different because the first two shifts change how people search.
This shift changes who is doing the search. It's not your customer browsing your website anymore. It's their AI agent, scan your site, pulling your data, comparing you against 30 competitors in a second, and making a recommendation before a human ever sees your page.
This isn't coming in five years. ChatGPT already has shopping features. Perplexity is recommending products.
Google is building agents directly into Chrome. It's happening now. And here's where it gets really interesting because this shift doesn't change how people find you.
It rewrites the entire customer journey from scratch.
For twenty five years, the customer journey looked the same. Someone has a problem, they search Google, they click a few links, they browse some websites, they compare options, they make a decision.
Humans were doing the work at every single stage. That's ending. Here's what's happening instead.
Someone tells your a a agent, find me the best marketing agency for my ecommerce brand under 5,000 a month. The agent goes out, visits dozens of websites, post pricing, services, reviews, case studies, cross references that against brand mentions and third party data, comes back with three recommendations.
The human picks one, done. They never visited a single site. Think about what just happened there.
The entire funnel awareness consideration decision got compressed into one AI conversation. The website didn't need to convince a human.
It needed to convince an agent. And that agent doesn't care about your beautiful design, doesn't care about your brand colors, or your clever headline. It cares about one thing, can I extract clear structured information from this site that helps my human make a decision?
If yes, you make the shortlist. If no, you don't exist.
So the obvious question is, what exactly are the agents looking for? Because once you understand that, you can reverse engineer exactly what to do.
And this is the part that almost nobody is explaining clearly yet.
AI agents don't read your website the way a person does. A person scrolls, skims, gets a feel for your brand, and an agent scans your code looking for specific signals. Here's what they're evaluating.
First, structured data, schema markup, and JSON LD. This is a language AI agents speak.
It tells them exactly what your page is about, what you sell, what it costs, and who it's for. Without it, you're invisible. Second, content clarity.
Not clever marking copy, clear, machine readable information. What do you offer?
How does it work? What's the pricing? If the answers are buried in five paragraphs of brand storytelling, the agent skips you.
And here's something most people don't realize. The same things that make your site accessible to people with disabilities, ARIA tags, clean HTML structure, clear labels, those are now the exact same things that make your site visible to AI agents.
. Third, API compatibility. Agents aren't just reading text.
They're looking for connections they can plug into. If an AI can check your inventory, pull pricing, or book an appointment through a clean API, you win.
If it has to guess by scraping your homepage, you lose. Google just launched something called the universal commerce protocol. That lets the agent handle everything from discovery to checkout.
The sites that plug into this kind of thing first get the sales. Fourth, your reputation across the web. Agents don't just read your website.
They cross reference everything. Brand mentions, reviews, sentiment, citations, other content. If no one else on the internet is talking about you, the agent doesn't trust you.
Fifth, freshness. When was this information last updated? AI agents heavily favor current data.
If your pages look stale, you look unreliable. Now, here's what's wild about this.
Most of those signals, they're the same thing that helped you rank in traditional SEO, structured data, clear content, strong brand presence, fresh information. The difference is for AI agents, these aren't nice to haves. They're requirements.
Miss one, and you're out. But here's the good news. Most of your competitors have no idea this is happening.
The bar to stand out right now is shockingly low. You just have to be ready before they are, and that's exactly what I'm about to show you. But before I do, if you want my company to just help you get set up for this new world of AI agents and make sure your marketing is intact so you could be getting the sales, check us out at NP Digital where we help companies across all industries and all sizes even in different languages all across throughout the world, helping them not only optimize for traditional SEO, but also LLMs and even AI agents.
Now here are the five changes that make your website AI agent ready. Most of these, you can do in a week. Step one, add schema markup to every important page.
Product schema, service schema, FAQ schema, review schema, this is the single highest impacting you can do. It takes an afternoon, and it makes every AA agent on the Internet able to understand exactly what you offer.
Think of it like putting a name tag at a networking event, except the event has a thousand people and everyone without a name tag is invisible. Step two, rewrite your key pages for clarity, not cleverness.
Every page should answer four questions in a plain language. What do you sell?
Who's it for? What does it cost? How does it work?
Header, short paragraphs, tables where it makes sense, you're writing for a very smart, very literal reader with zero patience. Step three, open your site to agent interactions. If you sell products, make sure your inventory and pricing are accessible through clean data feeds or APIs.
If you sell services, make sure your scheduling and availability is easy for agent to query. The businesses that let agents plug in directly, check stock, compare specs, and even complete a purchase, those are the ones that get recommended first.
Step four, build your brand signal everywhere. Get mentioned in industry publications.
Earn reviews on relevant platforms. Create content other sites want to reference. A agents trust you more when the rest of the Internet vouches for you.
Step five, keep everything current. Add last updated dates to your pages.
Refresh your core content quarterly. Update your reviews and case studies. Stale content is a trust killer for agents.
Now, there's a sixth move, and this is one that creates real long term advantage. Create content that teaches AI who you are and what category you own.
Don't just say, we're a marketing agency. Say, we help SaaS companies scale from 1,000,000 to 10,000,000 through performance marketing. Create comparison pages.
Build category specific guides. Publish listicles that position you as authority in your niche if you really are the authority in your niche.
Don't lie. The specificity trains the AI. The more specific and consistent you are, the more AI agents associate your brand with that category.
And once you own that association, that's when the real compounding starts. Now all that is a playbook, and it works.
But there's something even more important than what to do, and that's when you do it.
Every major Internet shift has had a window, a period where the early movers locked in advantages that lasted years. The businesses that figured out SEO in 2003, some still rank on page one.
The brands that went mobile first in 2010, they captured market share they hold today. The Gentic Web has a window too, and it's smaller than any window that came before it.
And here's why. AI adoption is moving faster than anything we've seen. Gartner is predicting that by 2028, a massive percentage of both b to b and b to c transactions will involve AA agents in the decision.
What took mobile five years to reach mainstream adoption? AA agents will do it in twelve to eighteen months. That means the early mover window isn't three to five years.
It's measured in months. And once AI agents learn which brands to recommend in the category, those recommendations compound.
The more recommendations lead to more traffic. More traffic leads to more data.
More data leads to more authority. More authority leads to more recommendations. It's a , and the businesses that get on it first spin faster every single day.
The businesses that wait, they're fighting a uphill battle against competitors who are already the default recommendation. That's the real opportunity here. Not AI content tools, not chat GPT tricks.
The real play is making sure that when AI agents start doing the buying and they already are, they're buying from you. Most of your competitors won't watch this video. Most won't even take action even if they do.
That's your advantage. Use it. If you wanna see how AI search is already reshaping which businesses get found, watch this video next.
The Hook
The bait, then the rug-pull.
The shift from search engines to AI agents as the first contact point in the buying journey is not a five-year forecast -- it is already running in ChatGPT shopping, Perplexity recommendations, and Google agent integrations. This video makes the case that the entire customer funnel has already collapsed into a single automated conversation, and lays out a concrete six-step playbook for getting your website onto the shortlist before the flywheel locks out latecomers.
Named ideas worth stealing.
03:34list
Five Agent Signals
- Structured data / schema markup
- Content clarity (what/who/price/how)
- API compatibility
- Brand signals across the web
- Freshness
The five criteria AI agents use to evaluate and shortlist websites without a human ever visiting them.
Steal forwebsite audit checklist, SEO pitch deck, content strategy brief
06:37list
Six Agent-Ready Steps
- Add schema markup to every important page
- Rewrite key pages for clarity (4 plain-language questions)
- Open APIs for agent interactions
- Build brand signals everywhere
- Keep everything current
- Create specificity-rich category-ownership content
The six-step playbook for making a website agent-readable, most achievable within a week.
Steal forclient onboarding checklist, SEO audit framework, web redesign brief
10:10model
Recommendation Flywheel
Agent recommendations drive traffic; traffic builds data and authority; authority earns more agent recommendations. Self-reinforcing loop favoring first movers.
Steal forgrowth strategy narrative, investor pitch, marketing ROI argument
How they asked for the click.
VERBAL ASK
06:02product
“check us out at NP Digital where we help companies across all industries and all sizes”
Mid-video sponsor block lasting roughly 35 seconds. Soft-sell framing as logical extension of the advice. Second CTA at 11:01 pointing to a follow-on video on AI search.
FROM THE DESCRIPTION
PRIMARY CTAWhere the creator wants you to go next.
Visual structure at a glance.
