In March 2025, I started an experiment: Could I get a business to rank #1 in ChatGPT and other AI search engines in its niche?
At the time, everyone was talking about AI search as “the future.” But nobody seemed to know how to actually optimise for it. There were theories, but no proven methodology.
Over the next nine months, working with a national medical company, I tested every possible theory I had heard about.
The short version: The result? #1 in ChatGPT and Gemini for their primary niche. Consistent recommendations across AI platforms. And a repeatable methodology that actually worked.
Three months ago, I joined Phantom Digital. One of my missions was to prove this system works for local businesses in any industry.
Today, when someone asks ChatGPT “Who’s the best foot and ankle surgeon in (the town of our client)?”, there’s one name that appears. When they ask Claude “Which accountant should I use in [Town of our client]?”, our client is the top recommendation.
This isn’t theory. This is a proven methodology, implemented in four different industries (so far), delivering consistent results.
This blog post describes exactly how AI search optimisation actually works. This is what we learned from nine months of systematic testing, and we are seeing the results now daily.
Why AI Search Matters Now
The way people search is fundamentally changing. Instead of clicking through ten blue links on Google, they’re asking ChatGPT, Claude, or Gemini for recommendations. People are getting either a single business as the answer, or a short list of businesses.
This is the zero click future. Users get their answer directly. No website visit required. True, this type of search is in its infancy for local businesses. Probably 90% of users still use Google for local search, and only around 10% use AI. But that 10% is growing month by month.
The opportunity: Most businesses aren’t optimised for this yet. The ones that are have a massive advantage.
The stakes: If you’re not in the AI answer, you don’t exist on that platform.
How AI Decides Which Businesses to Recommend
Through nine months of systematic testing, I’ve identified two layers of factors that influence AI recommendations. In many ways AI Search works a bit like conventional SEO. There are massive similarities.
However, there are also quite a few key differences
In a nutshell, the two ‘layers’ that you need are the following:
Layer 1: The Digital Foundation
The following four areas are more like conventional SEO, though they sometimes work slightly differently. These are the digital foundation that the AI needs to trust you as a business to mention.
1. NAP Consistency
Inconsistencies in Name, Address, and Phone number act as a red flag. If your website says one thing and Yelp says another, the AI may skip your business entirely to avoid giving users incorrect information.
AI doesn’t like risk. If things look a bit dodgy, it doesn’t like it!
2. External Citations
AI models are trained on massive datasets. If your business is mentioned in local news, industry publications, or “Best of [City]” lists, it strengthens the AI’s confidence that you’re a legitimate authority.
3. Local Entity Mapping
Using entity based SEO (mentioning local landmarks, neighbourhood names, and specific local services) helps the AI map your business to a geographic knowledge graph. The AI needs to understand you’re genuinely local, not just claiming to be.
This really is the same as conventional SEO, so no surprises here! But being local is important for local searches.
4. Review Sentiment
AI doesn’t just look at star ratings; it reads the text. If reviews frequently mention “fast emergency plumbing” or “excellent foot surgery outcomes,” the AI uses those specific phrases to justify recommending you for those queries.
These factors create the baseline credibility AI models need. Without them, you’re invisible. But they’re just the starting point.
Layer 2: AI Specific Optimisation
Beyond the basics, AI models prioritise businesses that make their data machine-readable and conversational.
A. Structured Data That AI Can Read
Schema markup tells AI models exactly what your business does, where you’re located, and what questions you can answer in a format they can ingest instantly.
Key Schema types:
- LocalBusiness Schema (hours, price range, coordinates)
- FAQ Schema (structured questions and answers)
- Review Schema (sentiment analysis in structured format)
While the principles are straightforward, the implementation requires careful technical work.
B. Conversational Content Architecture
AI models are trained on dialogue. Sites that use a Q&A format are much easier for AI to extract and present to users.
Instead of just a “Services” page, include sections like:
- “How much does [service] cost in [location]?”
- “What should I look for in a [service provider]?”
- “How quickly can [service] be completed?”
The content needs to mirror how people actually ask questions.
C. Freshness Signals
Traditional SEO can coast on old authority. AI models, however, increasingly integrate real time data. They prioritise businesses that show recent activity:
- New Google Business Profile posts
- Reviews from the last 30 days
- Updated holiday hours or seasonal changes
D. Trusted Directory Presence
ChatGPT and Gemini often cite aggregator sites to verify their recommendations. Being on Yelp, industry-specific directories, and local business listings is vital. These are “trusted seeds” the AI uses to validate its answers.
Case Study: The Foundation Project
National Medical Company (9 Month Testing Period)
Before joining Phantom Digital, I spent nine months developing and testing AI search optimisation strategies with a national medical company in a highly competitive space.
The Challenge:
- National-level competition
- Complex medical terminology
- Established competitors with strong SEO
- Zero AI search visibility initially
The Approach: Systematic testing across ChatGPT, Claude, and Gemini. Development of medical-specific structured data approaches. Creation of conversational content architecture around patient questions. Integration with medical authority signals.
The Results:
- #1 recommendation in ChatGPT and Gemini for their primary niche
- Consistent mentions across AI platforms
- Methodology validated through hundreds of test queries
The Insight: This project taught me something crucial: AI search optimisation isn’t about gaming an algorithm. It’s about making your expertise so clear, so structured, and so accessible that AI models have no choice but to recommend you as the authority.
The medical sector is notoriously difficult for digital visibility due to YMYL (Your Money, Your Life) standards. If the methodology worked there, I knew it would work in less regulated industries.
Case Studies: Local Business Success
When I joined Phantom Digital three months ago, I saw an opportunity: most local businesses had no idea this shift was happening. If we could apply these same principles to local service businesses, we could give our clients a massive competitive advantage.
We selected two test cases in different industries to prove the concept.
Case Study 1: Foot & Ankle Surgeon, Liverpool
The Challenge: A highly skilled surgeon in a competitive medical market. Multiple established competitors. Zero visibility in AI search results initially.
Our Approach: Using the methodology refined over nine months of national testing, we implemented AI search optimisation specifically adapted for a local medical practice:
- Optimised digital foundation (NAP consistency, citations, review management)
- Implemented AI-readable structured data
- Created conversational content architecture around patient questions
- Built local entity relevance through strategic content
The Results:
- Now appears as the primary recommendation in ChatGPT, Claude, and Gemini for Liverpool foot and ankle surgery
- AI models cite specific expertise areas when recommending
- Increased enquiry volume from patients mentioning “I found you through ChatGPT”
What Made the Difference: The combination of structured data, conversational content, and strategic local entity mapping created a clear “signal” that AI models could confidently recommend. While the principles are straightforward, the implementation requires careful technical and strategic work.
Case Study 2: Local Accountancy Firm
The Challenge: Established firm competing against national brands and local competitors in a midsize town.
Our Approach: Similar methodology adapted for professional services:
- Focus on service-specific questions accounting clients actually ask
- Integration of local business knowledge and community presence
- Strategic structured data implementation
The Results:
- #1 recommendation across ChatGPT, Claude, and Gemini for accounting services in their town
- AI models cite specific service offerings and local expertise
- Positioning as “the local expert” reinforced by AI recommendations
The Pattern: Both clients share common factors: comprehensive structured data, conversational content that answers real questions, and strong local entity signals. But the implementation details differ significantly by industry and competitive landscape.
The AI Search Audit Framework
Want to know if your business is positioned for AI search? Start with what I call the “Zero-Click Test.”
The Zero-Click Test
- Open ChatGPT, Claude, or Gemini
- Ask: “Who is the best [your niche] in [your city] and why?”
- Analyse the response:
- Are you mentioned?
- What justification does the AI give?
- Where is it pulling information from?
- Which competitors appear instead?
This simple test reveals whether your business is “AI-visible.” But understanding WHY you’re included (or not) requires deeper analysis.
Questions to Ask:
- What information is the AI citing?
- Where is that information coming from?
- What’s missing from your digital presence?
- What signals are your competitors sending that you’re not?
Based on nine months of testing, I’ve identified 47 specific factors that influence AI recommendations across these categories. It’s the same audit framework I used for the medical company and our Phantom Digital clients before implementing optimisation strategies.
What This Means for Your Business
The Opportunity Window:
AI search optimisation is still early. Most agencies aren’t doing this yet. Early movers have a significant advantage. The gap will close as more businesses catch on.
The Challenge:
Optimising for AI search requires:
- Technical implementation (Schema, structured data)
- Strategic content architecture
- Industry-specific adaptation
- Ongoing maintenance and refinement
The Timeline:
Based on my testing, getting from zero visibility to #1 recommendations takes 3 to 6 months when you know what you’re doing. It took me nine months to figure out what works. It took three months to implement for our Phantom Digital clients once we had the methodology.
The Future:
Within 24-36 months, AI search visibility may be as important as traditional Google rankings. The businesses that move now will establish authority that’s difficult for competitors to overcome later.
Next Steps
If you’re wondering whether your business is positioned for AI search, start with the Zero-Click Test. If you want to go further, but are not sure what to do, then reach out to us. We offer a free initial consultation where we can discuss you’re business and give you next steps.
At Phantom Digital, we’ve spent the last three months proving that the methodology refined over nine months of national testing works for local businesses in any industry. The results speak for themselves.
Want to know where your business stands? We’ve developed a comprehensive AI Search Audit that evaluates the 47 factors I discovered during testing. It’s the same audit we used before getting our clients to #1.
Contact Phantom Digital for Your Free AI Search Consultation.
About the Author: Martin joined Phantom Digital as SEO Lead after spending nine months developing AI search optimisation strategies in the medical sector. He specialises in helping local businesses establish authority in AI powered search engines.
