Free AI SEO Course: How to Get Your Brand Recommended in AI Search (AEO and GEO)

At MozCon London in 2025, our CEO Charlie Marchant asked a room of around 300 SEOs and marketers who had actively worked to get their brand showing up in AI tools. Almost everyone in that room used ChatGPT or Claude for their job every day. Fewer than 3% raised a hand.

That gap is the opportunity. AI Search platforms now shape which brands buyers shortlist, often before anyone visits a website, and most businesses still aren’t doing anything about it.

This free AI SEO course is built for CMOs, marketing managers, and business owners who want to decide what to do about AI Search, brief their team or agency, and prove it’s paying off. It’s the process our team runs for clients, with real examples throughout. We use AEO, GEO and AI Search Optimisation to mean the same thing.

This course is sponsored by Peec AI*, the AI Search visibility tool we use to track our clients.

Are You Showing Up in AI Search?

See how often ChatGPT, Gemini and Google’s AI answers recommend you, measured against your top three competitors, plus the three things we’d do first.

Get My AI Search Scorecard

Module 1: How AI Decides Which Brands to Recommend

Every Answer Draws on Memory and Web Search

When you ask ChatGPT for the best CRM for a small business, the brands in the answer come from two places.

Memory is what the AI model learned in training. Models learn by reading huge amounts of text and predicting the next word, so they don’t store a list of brands. They learn which words and ideas tend to appear together. A brand gets remembered when lots of independent sources mention it consistently alongside a topic. Ask for a good value skincare brand, and the model can answer from memory because it has seen The Ordinary described as affordable and science-backed so many times.

Memory has a cutoff date, so it misses new products, price changes and repositioning. You also can’t submit your brand to it. It’s built from what the rest of the internet says about you.

Web search fills the gaps. For anything recent, local, commercial or comparative (“best”, “vs”, “under £40”, “near me”), AI tools run their own searches in the background.

Tim’s question for your team: “If an AI described us from memory alone, what would it say, and is it still true?”

The Searches Behind the Search

When an AI tool searches, it breaks your prompt into several smaller queries and runs them in parallel. This is called query fan-out. Ask for the best shampoo for fine, frizzy hair, and it might search for ingredients that suit fine hair, lightweight shampoos, then reviews and prices of the brands it found.

To be recommended in the answer, your brand needs to show up in those background searches, and in the right context.

Recommended Without Being cited, and Cited Without Being Recommended

Tim ran a live search in Perplexity’s free plan for the best hyaluronic acid serum on a budget. It recommended The Ordinary, but every source it used was a third-party site: The New York Times’ Wirecutter, Fortune, NBC News, Health and People. None was a brand’s own website.

So your brand can be recommended without your website being cited. The reverse happens too: AI tools can quote your content without recommending you, which means you’re doing the work and a competitor gets the sale.

“AI reads many, many pages, often ones that you’d never rank for, and it turns all of this into one answer,” Tim Cameron-Kitchen, Founder of Exposure Ninja.

Why No Two AI Answers are the Same

AI answers are written from scratch every time, and they’re personalised to the person asking. In our own State of AI Search 2026 research, ChatGPT kept the same set of sources on consecutive days only 2% of the time (59 UK prompts, 24/06/2026 to 07/08/2026, 7,965 responses).

There’s no fixed “position one” in AI Search. Instead, you measure how often you appear across many prompts, platforms and days.

Module 2: The AI Search Opportunity

Many Searches Become One Conversation

Before AI, a buyer might Google a product, read a few articles, research ingredients, compare brands and read reviews, all across separate searches. Now that can happen in one conversation with ChatGPT, and the first time a brand knows that person exists is when they land on the website ready to buy.

This isn’t only happening in consumer markets. In G2’s March 2026 survey of 1,076 B2B software buyers, 51% now start their research with an AI chatbot more often than with Google, up from 29% the year before. 69% chose a different vendor than they’d planned to, and one in three bought from a vendor they’d never heard of before.

AI Traffic is Small, But it Converts

Adobe found that traffic from AI sources to US retail sites grew 393% year on year in January to March 2026, and that AI traffic converted 42% better than other traffic in March 2026.

Tim showed one of our lead generation clients’ analytics. AI referrals sent a fraction of the traffic paid search did, but converted at 32.1%. The more interesting number was organic search: its conversion rate had gone from about 2.5% to over 6% as the brand’s AI visibility grew. People were seeing the brand in ChatGPT, then searching for it on Google and converting.

Your Next Customer Might be an AI Agent

AI labs are building agents that research and buy on someone’s behalf. Bain forecasts agentic commerce could make up roughly 15% to 25% of US ecommerce by 2030.

Charlie tested this by asking an AI agent to buy coffee. It hit error pages on one site and went straight to a competitor instead. A human might persevere; an agent won’t.

The Six Metrics That Show You’re Winning

Traditional SEO had three numbers: rankings, traffic and conversions. AI Search adds new metrics to track:

• Mentions: an AI answer names your brand, with or without a link.
• Citations: an AI answer uses one of your pages as a source.
• Recommendations: you’re on the shortlist, the most valuable kind of mention.
• Visibility: how consistently you appear across the prompts you track.
• Share of Voice (SoV): how often you appear against your competitors.
• Sentiment: whether AI describes you positively, neutrally or negatively.

Module 3: How to Audit Your AI Visibility

Why Checking ChatGPT Yourself Doesn’t Work

Ask ChatGPT who the best company in your industry is, and it may well say yours, because it knows you work there. Answers also change every day and every time you ask, and buyers write long, conversational prompts you’d never think to test. It’s better to use an AI visibility tracking tool, not manual checks.

Four Questions to Set Your Baseline

1. Where are you visible today?
2. Where are you missing?
3. Who is appearing instead?
4. Which gaps matter commercially?

On the third question: in AI Search, your competitor is whoever the AI recommends instead of you. That might be a software tool if you sell a service, or even a government website, not just the businesses your sales team would name.

Setting Up Tracking in Peec AI

Tim walks through setting up a project in Peec AI* using Rocket Mortgage as the demo brand: add your brand and competitors, choose your prompts, select every AI platform, then let the data build for at least a week before calling it your baseline.

From there, Peec shows:

• Visibility, share of voice, sentiment and average position over time, filtered by platform.
• Perception: The market AI thinks you play in, the objections it raises about you (for Rocket Mortgage, things like high fees and closing delays) and fact-checks where AI contradicts your own website.
• Sources: The sites cited in answers that mention competitors but not you. That’s a ready-made digital PR outreach list, once you filter out competitors’ own sites.
• Actions: A ranked to-do list, split into work on your own site and coverage you need to earn.

Some of what comes back goes beyond marketing. Objections like poor communication or billing problems are customer research worth passing to product and customer success teams.

For a step-by-step version, see how to do an AI Search Optimisation audit.

Prioritising the Gaps Worth Closing

Focus on prompts where three things overlap: commercial intent, a competitive gap (competitors appear consistently and you don’t) and relevance (you have the product, service or expertise to deserve the mention). That’s where the extra revenue is quickest to win.

Putting a Number on the Gap

To turn this into a business case, use our AI invisibility cost calculator. Add the monthly searches behind your top buying questions, the share of your audience using AI, how often you’re missing, your enquiry rate, close rate and order value.

In Tim’s worked example, being missing from 85% of answers, with a 20% share judged winnable, came out at an estimated £3.1 million a year in extra revenue. It’s an estimate, not a forecast, but it lets a CMO say “closing this gap is worth X a year” instead of “our visibility score is 30%”.

Module 4: The AI Visibility Pyramid

AI Search Optimisation needs more than SEO. Traditional search has no memory, returns a page of links and has a page two. AI Search pieces together dozens of searches into one answer that might name two or three brands. As Tim puts it, “There’s no page two in AI Search.”

The AI visibility pyramid has three layers, built from the bottom up. First, though, decide what you want to be known for. The Ordinary’s positioning (affordable, transparent, science-backed) runs through its page titles, its About page and its reviews, which makes it easy for AI to describe. When Tim asked a large national company what made them different, the answer, after ten seconds of silence, was “it depends who you talk to”.

Layer 1: Technical Foundations

These are table stakes. Good foundations won’t win on their own, but weak ones will lose.

• Crawlability: AI platforms run separate training crawlers, search bots and user-triggered fetchers, and blocking one doesn’t block the others. Don’t block any of them. Check your CDN too: Tim points out that Cloudflare blocks AI training crawlers by default on new domains.
• Site Architecture: One hub per commercial topic, with sub-pages linked from it, no orphan pages and nothing buried more than about five clicks deep. Our episode on website architecture for AI Search goes deeper.
• Structured Data: The right schema markup for your page types, kept up to date.
• Fast, Working Pages: AI tools and agents won’t wait for a slow site. They move on to the next one.

Tim adds: “You’re not going to win with good technical SEO foundations, but you’re definitely going to lose without good technical foundations.”

Our technical SEO team can audit and fix all four.

Layer 2: On-site Content

Cover the questions behind the question. Peec’s fan-out view shows the queries AI tools actually run, such as “best digital mortgage lenders 2025”, comparisons and calculators. For one of our finance clients, it showed that people weren’t searching for the product at all. They were asking whether they should stop repayments on a type of loan, so we built content around that situation.

Answer the decision questions on your commercial pages: What exactly it is, who it’s for and not for, what problem it solves, how it works, what it costs, how it compares, what evidence backs it up and what its limits are.

Make it easy to extract. Tim compares AI tools to contestants on Supermarket Sweep: they grab what’s at the end of the aisle and move on. So answer the core question in the first few sentences, use clear subheadings, call out key data, and add summaries, FAQs and video. Our AI-optimised blog post template puts this into practice.

Give every page a clear next step, so a visit from AI turns into an enquiry or a sale.

Layer 3: Third-party Authority

This is the hardest layer and the one most brands skip. It covers three tiers:

• Editorial: Trade press, “best of” listicles and comparison pages. Being in a cited listicle matters, and so does where you rank in it.
• Reviews and Communities: Review sites, Reddit and forums. On Reddit, contribute, don’t promote. The community is good at spotting spam, which is why AI tools cite it so often.
• Your Footprint Beyond Your Website: YouTube, LinkedIn, podcasts and awards.

Client Example: ZUGU. ZUGU sells iPad cases, up against Amazon. We used digital PR to get it into the publications AI tools were already citing for iPad cases, including WIRED, ZDNet and MacRumors, alongside technical, on-page and content work. Across the campaign, ZUGU saw a 243% increase in AI platform traffic and exceeded its organic revenue goal by 132%. The campaign is shortlisted for Best AI Search Optimisation Campaign at the Global Search Awards.

If you want help earning this coverage, our digital PR team does it every day.

Module 5: How to Measure What AI Search Is Worth

AI visibility is the first step, not the goal: visibility leads to AI traffic, then leads or sales, then pipeline and revenue.

The hard part is that most AI influence never shows up as an AI click. People see your brand in an answer, then Google you or type your URL. We call this dark traffic. One of our finance clients’ analytics showed ChatGPT driving around 12 conversions a month, while prospects kept telling Sales they’d found the company by asking AI.

How to get closer to the real picture:

• Track AI referrals in GA4 under the AI assistant channel, and check Unassigned too.
• Watch branded search and direct traffic in Google Search Console and GA4 alongside your AI visibility.
• Watch your organic conversion rate. A rise as AI visibility grows is a sign AI is warming people up.
• Ask “How did you find us?” on forms and sales calls, and record it in your CRM. It’s the most reliable method of all. For Elite Renewables, one of our own businesses, ChatGPT is the most common answer on the phone, far ahead of what analytics shows.

The AI traffic you can measure is valuable. For one ecommerce client, ChatGPT visitors converted at 7.18% against 3.59% from Google, worth about $140 per visitor against $53. For a lead generation client, AI referrals converted at just under 14% against just under 6% from organic.

Then follow those leads all the way through: AI sessions, enquiries, MQLs, SQLs, customers and revenue.

We’re not going to get obsessed with our AI visibility number, because that metric isn’t going to fix anyone’s life. We’re going to get obsessed with how that metric turns into traffic, turns into conversions, turns into revenue.

Module 6: How to Run AI Search Every Week Without Growing Your Team

AI Search needs technical housekeeping, content refreshes, review generation, digital PR, monitoring, and reporting every week. Most marketing teams use AI tools individually, but generic prompts give generic output, so the marketing function itself hasn’t changed.

So we built Opal, the AI marketing system our team runs for clients and on the businesses we own. It works in four stages:

• Observe: It connects to analytics, Search Console, the CRM, the website and call recordings, and learns the business: goals, personas, tone of voice and offers.
• Plan: It works back from a leads or revenue goal to the levers that will move it, then to quarterly, monthly, weekly and daily tasks.
• Act: It does the work, such as publishing content, fixing pages and building redirects.
• Learn: It remembers feedback and results, so it gets better over time.

Every stage has human approvals, and our specialists direct it, check it and own the outcomes. Work only moves from “draft first” to automatic once we’re happy with it.

At Elite Renewables, one of our own businesses, Opal found a grant figure overstated by £5,000 across a 208-page sweep, rebuilt 124 pages lost across three site migrations with 86 redirects, and published a data piece on plug-in solar panels on the day they became legal. Elite has gone from around 25 leads a week to around 85, with a marketing team of one.

Five Things to Do Next

1. Establish your baseline. Track the prompts your buyers ask for a week and see where you stand against competitors.
2. Find the gaps worth closing. Prioritise commercially relevant prompts where competitors appear and you don’t.
3. Strengthen your pyramid. Fix the weakest of your three layers first.
4. Connect visibility to revenue. Set up tracking, ask how people found you and follow AI leads through to revenue.
5. Measure, learn and repeat, every week.

AI Search Optimisation isn’t about tricking ChatGPT. It’s about making your brand easy to find, easy to understand, easy to trust and easy to choose, then proving it turns into leads, sales and revenue.

If you’d like help putting this into practice, our AI Search Optimisation team does this for brands every day. Book a consultation and one of our AI Search specialists will look at your visibility across the platforms and tell you what we’d prioritise.

*Disclosure: This page contains referral links. If you purchase through these links, we may receive a commission at no additional cost to you.