The Match Protocol Score™ is a 0–100 measure of whether AI engines recommend a business. It runs blind buyer conversations on ChatGPT, Google AI Overviews, Perplexity and Gemini, reads the business's website the way an AI crawler does, and grades five stages: Found (20 percent), Matched (20), Recommended (35), Trusted (15) and Chosen (10). The overall score is capped by the AI Recommendation Rate, so a business AI never names cannot score above 49. It is the measurement layer of Match Protocol™, owned by Modern Asset Lab.
A buyer's AI conversation moves through these stages in order. Fail early and nothing downstream matters. That is why the stages are scored separately and why the lowest one is the one to fix.
Stage
The question
Weight
What a strong score looks like
Found
Can AI read you?
20%
Every AI crawler allowed, content visible without JavaScript, complete entity schema, llms.txt present.
Matched
Does AI name you for the right buyer questions?
20%
You are named in most conversations where a buyer in your market asks for what you sell.
Recommended
Does AI recommend you?
35%
AI picks you, not just lists you, in more than half of those conversations, across all four engines.
Trusted
Can AI verify you?
15%
Consistent, corroborated profiles and reviews; your site cited as a source in the answer.
Chosen
Is your site built to convert the human?
10%
Clear offer, proof, a next step, and a page that answers the question the buyer arrived with.
A score should never be able to lie. Without a cap, a business with perfect schema, a fast site and glowing reviews could score in the 70s while AI recommends its competitor in every single conversation. That business would feel fine and keep losing.
So the Match Protocol Score™ is capped by the AI Recommendation Rate: the share of blind buyer conversations in which AI actively recommended the business. Never recommended means the score cannot leave the Ghosted tier, whatever the other four stages say. The cap lifts as the recommendation rate rises.
This is also why Recommended carries 35 percent of the weight on its own. It is the stage that produces customers. The other four exist to make it happen.
A dental practice scores Found 90, Matched 60, Trusted 85, Chosen 70. Strong on paper. But across 32 conversations, AI recommended it twice: a 6 percent recommendation rate, Recommended stage 12.
Weighted: (90×0.20) + (60×0.20) + (12×0.35) + (85×0.15) + (70×0.10) = 53.9. That is a Ghosted-tier score, and because AI recommended the practice in only 2 of 32 conversations, the recommendation cap keeps it there whatever the other four stages say.
The lowest stage is Recommended. The layer that moves it is Targeted Extraction: answers written to be lifted into an AI response, and FAQ content for the questions buyers actually ask. That is the first month's work.
The tier names are blunt on purpose, and the ranges follow the U.S. school grading scale, so anything under 60 reads as a fail. AI search is a matchmaker, and most businesses are not getting the introduction.
AI rarely or never recommends you. Eight in ten businesses in the 2026 Benchmark landed here, including many with perfectly good websites. The work starts with whichever stage is lowest, and it usually pays back fastest.
AI knows you and names you, but rarely picks you. You are on the shortlist and losing the final call. Corroboration and extractable answers are usually what separates you from the business being chosen.
AI recommends you in a meaningful share of conversations. Few businesses reach this tier: 5 of 100 in the benchmark. The remaining gap is usually one engine that has not caught up yet.
AI recommends you in most conversations where you fit. Only 2 of 100 businesses in the benchmark reached this tier. What remains is usually one engine still to win, or a handoff that loses the click.
AI recommends you in most conversations where you fit, across all four engines, and the buyer who clicks is met by a page that closes. No business in the benchmark reached it; the highest score was 88. That is the opportunity.
100 U.S. small businesses, October 2026. Use these to read your own stage scores.
80 of 100. AI can read most businesses. Being readable is table stakes, not an advantage.
Matched42 of 100. Named in some conversations, rarely in most. 31 percent were never named at all.
Recommended18 of 100. 36 percent were never recommended. 62 were recommended three times or fewer.
Trusted88 of 100. AI can verify most businesses exist. It still will not pick them.
Chosen74 of 100. Most sites can take the handoff. The problem is upstream: the buyer is rarely sent.
The Match Protocol Score™ asks ChatGPT, Google AI Overviews, Perplexity and Gemini the questions buyers ask, never naming the business, and weights five stages — Found 20 percent, Matched 20, Recommended 35, Trusted 15 and Chosen 10 — into one 0–100 score, capped by how often AI recommends the business. Recommended carries the most weight, so a business AI never recommends cannot leave the Ghosted tier.
The AI Recommendation Rate is the share of blind buyer conversations in which an AI engine actively recommended the business, as opposed to merely mentioning it. It is reported per engine and overall, it drives the Recommended stage, and it caps the overall Match Protocol Score™.
A mention is not a recommendation. Many businesses are named in a list of options and never chosen. The Score counts mentions under Matched and recommendations under Recommended, weights the recommendation three times heavier, and never asks about the business by name, which would bias the answer.
So the Score cannot look healthy while AI sends every buyer to a competitor. The cap ties the overall number to the AI Recommendation Rate, the only stage that produces customers. As the recommendation rate rises, the cap lifts.
Monthly during active Match Protocol™ work, because engines re-crawl and re-weigh on their own schedules and corroboration takes time to propagate. Otherwise quarterly, and after any major change to your website, profiles or offer.
The assessment at match.modernassetlab.com runs the lite version of the methodology for any business. Businesses that qualify for a Match Protocol™ engagement receive the full four-engine scan as part of their Strategy Session.
The Match Protocol Score™ was developed by Clint Rogers and is owned by Modern Asset Lab, a Tulsa, Oklahoma client acquisition agency. It is the measurement layer of Match Protocol™, which was first implemented on Creativo AI, the Techstars-backed company Rogers leads. The same methodology produced the 2026 Match Protocol™ Benchmark.
Ninety seconds. Four engines. One honest number, and the first thing to fix.
The strategies discussed herein are business growth systems. This is not financial advice. Individual results vary based on implementation, market, and execution. Modern Asset Lab is a division of Do More Good Inc.