The shelf is built. Here is what is on it.
What we learn doing AI visibility work for medical, legal, wealth, and mental health practices: what moved a number and what did not. Three pieces are up. Every figure names its source with a link, and we note when a survey came from a company with something to sell.
Three pieces, published.
Any order works. The first is the argument the other two sit inside.
People bring big questions to a machine they say they can't verify, so the engine hands over a name. What it takes to be that name.
The free audit, the alarming score, the jargon, the monthly retainer. One question sorts the companies selling the work from the ones selling the invoice, us included.
Most fixes need doing once. Sort by how often each item recurs and you'll see what's worth paying for monthly.
Four rules, decided before the first piece.
Most writing in this field fills a content calendar. We'd rather publish less and be quotable.
Client projects (with permission, identifying details removed), changes to our own site, and probes we ran across engines. No roundups of other people's takes.
Everyone publishes what worked. We print the attempts that went flat next to the ones that moved.
Every rate comes with its run count, market, engines, and time window. Without those, a number is decoration.
A physician or attorney at the end of a long day should see what it means for their practice. Jargon gets defined.
How the writing is organized.
Newest first, dated. Two shelves are empty and stay labeled that way until something real goes on them.
Run counts, frozen question sets, mentioned versus cited, confidence and its limits, and where our own method is weaker than we'd like.
How a dermatology patient and a bankruptcy client ask AI differently, which sources the engines lean on in each field, and what moved. Identifying details removed.
What the engines appear to read and reward, judged from the outside. Nobody outside a model lab knows the internals, and these pieces say so.
Board and bar advertising rules meet AI answers: testimonials, superlatives, outcome claims. The most compliant marketing is usually what engines trust most.
What agencies sell, how pricing is built, and which work recurs versus only the invoice. We sell this work too, so we get the same test.
On the desk, not yet on the shelf.
In progress, so nothing to click yet. Listing them keeps us honest about what is finished.
How wide the confidence band is at five runs, at twenty, at a hundred, and when more runs stop being worth the money for a local practice.
Why one old negative review keeps surfacing when someone asks an engine about a practice by name, across engines, and what an owner can do about it.
Our working list per vertical, how we built it, and how often it changes. The useful list is short, whatever the fifty-directory submission services say.
One question asked from three cities, answers side by side. The case for geo-varied probing, made with results instead of adjectives.
No dates promised. A piece publishes when the work is finished and checkable. To hear when the next ones land, write to human@aeoptim.com and we'll add you to a list used for nothing else.
Three more things worth your time.
Three pieces is not a library yet. Meanwhile, these pages show how we think.
A live engine asked your clients' question, answer returned word for word. No card, no call.
Run it →Our measurement method in full: how we probe, the runs behind each rate, mentioned versus cited, and the method's limits.
Read the method →Your clients' questions, answered on your own pages in the format engines quote, with paste-ready schema. The fastest paid step from knowing to fixed.
See the kit →The most useful thing on this site is still free.
One live answer will tell you more about your practice than anything published here.
Run your free check →Want a heads up when the next pieces land? human@aeoptim.com