What AI already says about you.
This is Edition 08 of Reputation means business. Everyone is asking what to build with AI. Few are asking what AI already says about them. Laure tested that question on herself and Olympia Advisory, comparing what several AI tools returned.
Many of us are spending a great deal of time asking what we could or should create with AI. There is another question receiving far less attention than it should: what does AI already say about us?
The experiment.
I ran the experiment on myself and Olympia Advisory, putting the same question to several different AI tools: what was I known for professionally, how was the business positioned, and which evidence supported the answer.
Every answer found the material I had built, and independent proof beyond it too. The answers were broadly consistent, but not identical.
One was thorough and careful enough not to say anything without the evidence. Another returned detailed results but assumed a specialty I have never claimed. A third pulled very little information.
What you can actually control.
AI selects from what is available and fills gaps with assumptions whether they are true, accurate, or not. While you cannot control the precise answer the tools produce, you can influence the evidence from which that answer is assembled.
A thin or generic answer usually means the evidence does not communicate your work clearly in the right places. Publishing more will not necessarily fix that. Consistency and selectivity across more than one independent source are more likely to carry weight than sheer volume of self-published material.
The objective is not to make every AI tool produce identical wording, which would not be possible anyway, but to make sure the same association holds up when the source is not just one channel.
One channel is not enough.
LinkedIn is also one of the harder places for a tool to see clearly: it restricts scraping and crawling more than the open web does, so what a tool pulls from it is often a partial picture rather than the current one.
What actually holds an association together across different tools is third-party confirmation: a mention, a reference, or a piece of coverage that came from somewhere other than you, telling the tool the claim was not simply self-reported.
Clients, candidates, journalists, recruiters, and conference organisers are increasingly using AI to prepare for meetings, and the answer they get may be someone's first introduction to you, before you know the research happened.
What does yours say?
On 24 September, I will unpack this live in the next Olympia Advisory Briefing, What AI already says about you, and I would like you there for it.
Try the experiment yourself, and let me know whether your answer is what you expected.
