Last Sunday we ran an experiment on ourselves. We asked ChatGPT and Perplexity five questions an Edmonton business owner might actually type: who builds websites for contractors in Edmonton, best web design company for trades businesses, that kind of thing. Ten runs total, fresh sessions, logged out, everything recorded word for word.
MM8 Studio appeared in zero of the ten answers.
Most agencies would bury that result. We are publishing it, because it is the honest starting line of something we intend to document in public: what it actually takes for a local business to get recommended by AI tools, measured monthly, with the receipts kept.
Why this matters to your business, not just ours
A growing share of your customers no longer scroll through ten blue links. They ask a question and get an answer with three or four business names in it. If your company is one of those names, you get the call. If it is not, you do not even know the conversation happened. There is no page two to claw back from. You are either in the answer or you are invisible.
When we ran our experiment, the tools confidently recommended a handful of established local agencies. Were those the best choices for a trades business? In some cases the recommendations were reasonable. In others, the reasoning the tools gave was thin or outdated. But it did not matter, because the person asking never sees what got left out.
What actually gets a business into the answer
Nobody outside these companies knows the exact recipe, and anyone who claims certainty is selling something. But the pattern across what the tools cite is consistent enough to act on, and it is refreshingly boring.
Say plainly what you do and where. Not “innovative solutions for your space,” but “we hang and finish drywall in Edmonton and St. Albert.” AI tools recommend businesses they can describe. If your own website cannot state your trade and your service area in one sentence, no machine will do it for you.
Answer real questions on your site. The tools lift from pages that already read like answers. What does a ceiling repair cost, how fast can you come out, do you handle insurance work. A page of genuine answers to genuine questions outperforms a page of slogans every time.
Keep your details consistent everywhere. Your name, address, and phone number should match on your website, your Google profile, and every directory that lists you. Machines cross check, and inconsistency reads as unreliability.
Structure the page so machines can read it. Behind the scenes, pages can carry structured data that labels what each piece of content is: this is a question, this is its answer, this is a business, here is its service area. Invisible to visitors, load bearing for AI tools. This is technical work, and it is a real part of what separates a site that gets cited from one that gets skipped.
Stay visibly alive. A site that has not changed since it launched looks abandoned to a machine. Regular, genuinely useful publishing signals that the business behind the site is operating.
The experiment, ongoing
Here is our commitment. We have rebuilt our own site around exactly the principles above, the same ones we install for clients. Every month, on the 19th, we rerun the same ten questions the same way and record what changed. Whatever happens, we publish it: the month we first get mentioned, the months nothing moves, all of it.
If it works, you will be reading the most honest case study in Edmonton marketing. If it takes longer than we would like, you will be reading that too. Either way, you will know it is real, because you watched the starting line get drawn.
Zero out of ten. That is where this starts. Check back.