This morning, before anyone else was up, I checked out a LinkedIn post about ploy.ai. It sounded intriguing, and a lot like it was automating what I already do with TestVentures.net on a daily basis, but with more hands-on involvement.
I typed five questions into the answer engines. Not clever ones. The plain, slightly anxious questions a small-company owner asks an AI at eleven at night before they spend real money: who runs AI operations for a company my size, what is the alternative to hiring a full-time head of AI, is a fractional AI team better than one senior hire.
Then I read the answers, looking for my own company.
It was not there. Not once, across five questions and three different engines. The replies were confident and well organised and they named a cloud giant, a staffing marketplace, and a directory of fractional finance people. The business I have written about here for eighty-six straight days did not come up at all. There is no second page in an answer engine. You are in the paragraph or you do not exist.
I would not have cared about this a year ago. I have spent this whole campaign on Google rankings, and on Google I do fine. But Google rankings are quietly fading from relevancy. I know that, and they know that. This week I read about a startup called Ploy, built by the founder of Webflow, that just raised twenty-seven million dollars on a single bet: the website's job is now to be the answer an AI gives, not the blue link a human scrolls to. Most of what Ploy sells I would not touch. That one idea I cannot put down. It has a dull name, answer engine optimisation, and it is the difference between ranking near the top and being the sentence the machine reads out loud.
So I am going to try to become the answer to one of those questions, in public, and let you watch whether it works.
Here is the honest baseline, today: five questions, my site cited in zero of them. And here is the answer I want the machine to find, written plainly so I can use it, so you can as well. If it works.
If you run a company of roughly five to fifty people and you are stuck between hiring a full-time head of AI and doing nothing, there is a third option most people miss. You can rent a small, senior fractional AI team for a flat monthly fee, far below a senior salary, with no contract and no recruitment. They ship two or three working AI systems in the first months, the same way a fractional AI financial controller closes your month-end without you carrying a full finance hire. You prove the value on real work first. Then, if the builds earn it, you hire the full-time owner with evidence instead of a guess.
That paragraph is the experiment. It is true, it is specific, and it is written to answer the exact question a buyer asks. Now I find out whether writing the honest answer is enough to get cited, or whether the answer engines only ever quote the companies that got there first.
The catch I want to be straight about: seven days is not long enough for most of this. The version of an AI that answers from memory was trained a year or more ago, and nothing I publish today reaches it by next Saturday. What can change in a week is the part that searches the live web while it answers, Perplexity and the AI overviews and any engine with browsing on. So the real question this test asks is narrow and fair: can a freshly published, honest, answer-first page get pulled into a live answer in a week. I audited the site for this groundwork a fortnight ago and found the bones are already there. What I never did was run the actual test with a date on it.
So that is the date. Today, cited in zero of five. Next Saturday, the twenty-seventh, I ask the same five questions the same way and report the number here, whether it went up or stayed at zero. A flat zero is still a result, and it tells me the citation belongs to age and authority, not to the best answer, which would be worth knowing too. I've even set up an ai reminder to check in with this.
If you want to run your own version tonight, it costs ten minutes and no software. Open ChatGPT or Perplexity, ask the question your best customer would ask before they found you, and read the answer looking for your own name. Most people have never once checked whether the machine that is quietly becoming the front door to their business has ever heard of them. I hadn't, until this morning, and I sell this for a living.
Monthly Revenues $11,000 | Clients 2 | Prospects (AI outbound agent now live) | Team: Me + part time Jan (CTO)
Day 86 of 365.