For years, when people wanted to know about you, they searched your name and looked through a page of links. That began to change after ChatGPT launched publicly in late 2022. Today, more people are asking a reasonable question: Why look through ten links when I can ask an AI and get a direct answer?
Sometimes that answer is accurate. Other times, it describes an event from 2011 as though it happened yesterday, leaves out something that would change your understanding, or draws a conclusion the facts do not support. It can sound equally certain either way.
Have you asked your favorite AI about yourself? If what came back surprised you, you already understand why this matters.
Much of my work now focuses on the relationship between Google search results and AI-generated answers. The two are connected, but they can tell very different stories about the same person. Why does a particular result appear? What information is an AI using, and what has it missed? When the picture is wrong, incomplete, or years out of date, what can actually be done to correct it? These are questions I work through with clients at RMG Digital Solutions, where we strengthen the public record that search engines and AI systems draw from. We publish our methodology and our standards because anyone considering this work should be able to examine how we do it.
Terms such as GEO and AEO have emerged to describe aspects of this work, and there seems to be a new acronym every week. I am more interested in what can be established: where the information comes from, how AI systems interpret it, and whether the account they give of a person is true.
Everything about how the firm actually works is set out at RMG rather than here: how an engagement runs, what it has looked like for clients, and the standard we hold ourselves to. I would rather you read it there, in full, than take my word for it on this page.
What I am working on
I registered the domain autoregressive.ai weeks after ChatGPT launched publicly, and the word has become the center of how I think about this. A system like ChatGPT writes one piece at a time, each shaped by everything that came before it. A reputation works the same way: what has already been said about you sets the odds on what gets said next, and the machines have now joined the conversation. The practical consequence is that the record has to be built deliberately, in the open, from true things, so that the next answer starts from a better place than the last one.
That is what I write about here when I have something worth saying. At the firm, it is the work of our Enhance practice: a record you have actually earned, made findable to people and machines alike. When something in the record is simply wrong, Restore is where we begin instead.
If you would like to know where you stand before deciding anything, the firm's AI Answer Audit puts ten questions to five AI systems and Google and gives you a dated record of the answers and a plan. It costs $2,500, credited in full if you go on to work with us within 90 days.
And if you are weighing firms, ours or anyone else's, I have written down the questions I would ask before hiring one.