Ranking in AI search is not a replacement for SEO — it’s roughly 70% the same work plus a distinct 30% that most investors are ignoring. Glen Petersen, head of partnerships at Bateman Collective and a former PPC director who managed about $11 million a month in Google spend, puts the numbers bluntly: around 2 billion AI searches happen per month against 8 to 13 billion Google searches per day. AI models also have no independent index. They scrape Google and Bing. If you rank poorly organically, ChatGPT and Gemini will never find you.
That means the entry point for investors who skipped SEO is the same entry point it always was — fix the site, build authority, publish real content — with a handful of new rules layered on top.
Below: what specifically changes for AI optimization, why prompting 1,000 city pages backfires, the month-by-month build order Petersen uses with clients, and whether the people searching this way actually match your buy box.
Key takeaways
- AI platforms have no independent index — they pull from Google and Bing results, so organic rankings are a prerequisite for AI citations, not an alternative to them.
- Backlinks carry less weight in AI optimization than unlinked brand mentions. Being talked about on Reddit (ChatGPT’s heavy source) or YouTube (Gemini’s) builds authority without a single link.
- Content under 300 words isn’t recognized as credible by AI systems. 400+ words works, 1,000+ is better, and Petersen’s target range is 1,500 to 2,500.
- Move FAQs off a standalone FAQ page and into each individual service page — that’s what gets matched to a conversational query.
- Expect 4 to 6 months before AI platforms start citing you and 9 to 12 months before rankings produce meaningful lead flow. Fix site speed, forms, Google Business Profile and schema in months 1-3 before scaling content.
From the Investor Fuel Show
This article draws on an interview with Glen Petersen of Bateman Collective on the Investor Fuel Show, hosted by Mike Hambright. Watch or listen to the full interview.
AI Search vs Google: The Volume Numbers That Should Set Your Priorities
Roughly 2 billion AI searches happen per month. Google handles 8 to 13 billion searches per day. Petersen pulled both figures, and the gap should settle the “PPC and SEO are dead” argument for anyone still having it.
What is true is that about 30 to 40% of searches now have some AI touchpoint — an overview, a summary, a chat handoff. That’s real and it’s growing every month. But it’s a change in how results get presented, not a replacement for the underlying search economy.
The structural point matters more than the volume. AI models do not maintain their own index of the web the way Google does. When you type a question into ChatGPT or Gemini, the model runs multiple searches across platforms like Google and Bing, then assembles an answer from what it finds. It scrapes the pages that rank.
So the investor hoping AI is a side door around a decade of SEO work has it backwards. As Petersen put it, without SEO, AI models will never pick you up. If you’re sitting on page 70 for “sell my house fast [your city],” you are invisible to the crawl that feeds the answer.
The flip side is that competition is thin. Petersen’s read on the investor space is that only 10 to 20% are doing the right things to rank organically, and below 10% are doing anything deliberate on the generative engine optimization side. That’s a low bar in a channel where the work compounds.
Treat AI visibility as a layer on top of organic performance, and set your budget accordingly — the foundation is still the foundation.
What Actually Changes: The 30% of AI Optimization That Isn’t SEO
Petersen’s research puts about 70% of AI optimization squarely in standard SEO territory. Both systems chase the same trust signals: authority, consistency, relevance, accurate information. The remaining 30% is where the tactics diverge, and it’s specific enough to act on this week.
Brand mentions replace backlinks. The paid-backlink era — a thousand dollars for ten thousand links — did real damage to the industry and AI systems largely ignore it. What builds authority now is being mentioned. An article that names your company without linking to it still counts. That’s a different outreach motion than link building.
Word count has a floor. Content under 300 words isn’t recognized as credible. Over 400 works. Over 1,000 is better. Petersen’s working target for client content is 1,500 to 2,500 words. Under traditional SEO you could stack hundreds of 300-word pages and see benefit. That approach is now dead weight.
FAQs move inside the page. The standalone FAQ page was built for a keyword-matching world. Put the FAQs directly inside each service or product page instead, so the question and the answer sit next to the context. A page that answers eight real questions about tax-lien sales is far more usable to an AI than an FAQ page with one line about it.
Different platforms lean on different sources. ChatGPT weights Reddit heavily — it’s where people discuss things in the open and where your brand name gets repeated without links. Gemini leans on YouTube, unsurprising given both are Google products.
Underneath all of it is intent-matching rather than keyword-matching. Search “small poodles” on Google and you get pages about small poodles. Ask an AI and it works out whether you want to buy one, see photos, or compare breeds, then routes you to the best provider for that intent. Write to the question behind the query, not the string.
Without SEO, AI models will never pick you up. Without having any kind of content online, without having the right stuff and ranking well, AI models will never pick you up.
— Glen Petersen, Bateman Collective
Why Mass-Produced AI Articles Backfire
The obvious play is to prompt ChatGPT for ten articles across a hundred DFW suburbs, get a thousand pages back in ten minutes, and dump them on the site. It doesn’t work, and it can actively cost you.
Google has started suppressing rankings when it detects AI-generated content. Petersen notes that detection has gotten specific — some of the tools college professors use track how a document was typed, including when large chunks were pasted in. The signal isn’t just linguistic anymore.
The workable middle ground is AI-assisted, not AI-generated. Have the model produce an outline or a rough summary, then write and expand it yourself, adding the expertise only an operator has. Petersen’s view: retyping and building on an AI draft is more ideal than pasting it, and content that comes out of your own head still works best.
On hyper-local pages, there’s a narrow tolerance. If you genuinely have similar expertise across a metro, swapping Dallas for Fort Worth to make a page relevant to that submarket is acceptable and won’t hurt much. What triggers problems is scale — a hundred pages identical except for the city name reads exactly like what it is, and Google treats it as copy-paste across the nation.
The practical rule: be comfortable with some duplication, not a lot. Each city page should carry something real — a neighborhood you actually buy in, a permit quirk, a price band, a deal you closed there. That’s the part AI can’t produce for you, and it’s the part that makes the page worth citing.
The Build Order: What to Fix Before You Write a Single Article
Content is not step one. Petersen’s timeline puts months one through three on the site itself, and most investors need that window.
- Months 1-3 — technical and foundational. Fix slow load times. Test every form, because a surprising number are silently broken. Rebuild pages that were written to inform into pages built to convert. Get a real Facebook presence live. Set up and fully configure your Google Business Profile — Petersen calls this a source of free leads that most operators leave half-finished. Implement schema, which almost always requires a developer.
- Months 3-6 — content at volume. Once the foundation holds, start publishing consistently. A few organic leads begin trickling in. AI platforms typically start picking you up in the 4 to 6 month range.
- Months 6-12 — ranking and correction. This is where lead flow becomes visible. It’s also where you review your AI citations and adjust. Petersen sees clients showing up in AI answers for topics that aren’t their actual expertise, which is a signal to redirect the content plan.
- 9-12 months — results. That’s when clients typically see meaningful rankings and lead volume.
Here’s why the early months feel like nothing is happening. Say you start on page 70 of Google and gain eight pages a month. By month five you’re on page 30 — enormous movement, zero leads, because nobody goes to page 30. It takes nine months just to reach page one, and page one isn’t the top three. Progress is real the entire time; it just doesn’t monetize until the last stretch.
Budget and expectations should be set against that curve. If you need leads this quarter, that’s a PPC or direct mail problem. This channel is the retirement account.
Supporting Content: Reviews, Video and the Local Citation Hack
Articles alone don’t carry authority. Petersen’s list of content supports is short and concrete: your own photos, your own video, first-hand knowledge in the text, and reviews from clients that reference the same things your articles cover.
The review tactic is the one most investors can implement immediately. When a seller leaves you a review, coach them to name their city and the specific service. “I needed to sell my home in Mesquite fast after inheriting it” does something a generic five-star review does not — it registers your business as hyper-local for that city and that service, so people searching in that city start seeing you.
This is free, it’s repeatable on every closing, and it compounds. Build the ask into your post-close process the same way you’d build in a referral ask. Send the seller a one-line suggestion of what to mention rather than hoping they get specific on their own.
Off-site placement matters for the same reason. Since ChatGPT weights Reddit and Gemini weights YouTube, content that lives outside your domain can produce citations your blog never will. A YouTube walkthrough of a rehab in a named neighborhood is doing double duty — it’s a video asset and a Google-owned source Gemini already trusts. Honest participation in relevant Reddit threads, where your brand gets named in discussion, builds the unlinked mentions that carry weight now.
None of this replaces the articles. It’s what makes the articles credible enough to get pulled into an answer.
Who Is Actually Searching AI — And Whether It Matches Your Buy Box
Before you reallocate budget, check whether your seller actually uses these tools. Petersen’s read on the AI search audience: majority under 45, concentrated in metropolitan areas, more educated on average.
Compare that to the typical investor buy box — lower income, outer-suburb but not fully rural, often older homeowners in distress. There’s overlap, but it isn’t a clean match today. An investor whose entire pipeline comes from probate leads on 60-plus sellers should be clear-eyed that this channel is not going to fill that pipeline next quarter.
The second qualifier is funnel position. People are not typing “I need to sell my home fast” into ChatGPT. They’re typing something like “the county just put a tax lien on the house I inherited, what are my options.” The AI returns possible solutions, one of which is selling to a cash buyer. That’s genuinely top of funnel, and it needs a nurture path behind it — the lead is not arriving ready to sign.
What makes it worth building anyway is direction of travel. Every monthly update Petersen sees shows AI search volume climbing, and demographic adoption widens as it goes. Getting positioned now means being visible when the curve catches your seller.
The proof point he cites: one Bateman client generated $500,000 in SEO-attributed revenue last year, with roughly 25% of it traceable to AI search. Two caveats he attached himself — that client had been doing SEO for years, and they’re in a large market. Tulsa will not produce Chicago or DFW numbers. The pattern is real; the magnitude scales with market size and prior investment.
Frequently asked questions
Is SEO still worth doing for real estate investors now that AI search exists?
Yes, and it’s now a prerequisite rather than an alternative. AI models don’t maintain their own web index — they run searches against Google and Bing and build answers from what ranks. Weak organic performance means AI systems never encounter your pages in the first place.
The volume gap also matters: roughly 2 billion AI searches a month against 8 to 13 billion Google searches a day. Run both together, since about 70% of the work overlaps anyway.
How long does it take to start showing up in ChatGPT and Gemini results?
AI platforms typically start picking up a site around 4 to 6 months into a serious effort, assuming the technical foundation is already fixed. Meaningful ranking results and lead flow land closer to 9 to 12 months.
The lag is structural. Moving from page 70 of Google to page 30 is major progress that produces zero leads, and at a realistic climb rate it takes about nine months just to reach page one.
Do backlinks still matter for AI search visibility?
They matter far less than they do for traditional SEO. AI systems build authority partly from unlinked brand mentions — an article, forum post or review that names your company without linking to it still contributes.
That shifts the work from link acquisition to earning conversation. Getting your brand named repeatedly on Reddit, which ChatGPT weights heavily, can move the needle faster than a link-building campaign.
How long should a blog post be to get picked up by AI search engines?
Aim for 1,500 to 2,500 words. Content under 300 words isn’t recognized as credible by AI systems at all, 400-plus is the working minimum, and over 1,000 is meaningfully better.
This is a real break from older SEO practice, where a stack of 300-word pages could still help you. Those pages now do nothing for AI visibility.
Can I just have ChatGPT write city pages for every suburb in my market?
No. Google suppresses rankings when it detects AI-generated content, and detection has gotten sophisticated enough to flag pasted content patterns, not just writing style.
Swapping one city name for another across a few pages is tolerable if you genuinely serve both markets. A hundred near-identical pages is not. Use AI for an outline or first draft, then add the local specifics — neighborhoods, price bands, deals you’ve actually closed — that only an operator in that market can supply.
The bottom line
Start with your website, not your content calendar. Run a technical audit this week — load speed, form submissions, whether your pages are built to convert or just to inform — and get your Google Business Profile fully configured before you commission a single article, because content published on a broken foundation compounds nothing.
