When Different AI Ecosystems Tell Different Stories About You

Your professional reputation is not what comes up as a single search result anymore. It varies depending on the AI system used, what it retrieves, and when you ask.

One Person, Different Answers

I asked a few different AI systems the same question about me after noticing how much AI summaries can vary. The answers were not identical. What was interesting is what each system chose to notice.

I have built my career across countries, industries, and professional communities, so I have become familiar with the idea that the same experience can mean different things depending on who is looking at it. Moving to New Zealand made that particularly tangible: experience that was well understood in one market sometimes needed context in another.

Now with AI, the interpreter is a machine. That adds another layer: what does a particular system decide to notice, prioritise, and tell someone about you?

Why the Difference?

Google AI Mode, ChatGPT Search, Perplexity and Claude combine retrieval, source selection and synthesis differently. They also differ in when and how they retrieve fresh information.

Here is a simplified view:

The AI systems change constantly, and much of the technology is proprietary. There is no single AI database containing a definitive version of you. Two systems can look at broadly the same professional footprint and still produce different answers.

AI Has Two Clocks

There is one more technical difference. Training and retrieval operate on different clocks.

The training clock governs what is set in a model's underlying knowledge. An LLM does not continuously rewrite that knowledge whenever something changes online.

The retrieval clock is different. When an AI system has web access, it can retrieve more current information at the time of the query rather than relying only on trained knowledge.

That creates a freshness gap, but the size of that gap depends on the AI system. Your professional reality can change today. Your source can be updated today. Whether the AI answer changes today depends on whether that system decided to look.

The Missing Links: What Happens Under the Hood

To manage the source layer effectively, we must understand how AI systems find and interpret information.

Knowledge graphs and structured data. Search engines and other systems use structured information to understand relationships between entities such as people, organisations, roles and publications. Clear, consistent structured data help establish those relationships, although it does not guarantee what an AI system will finally summarise.

Semantic search. Retrieval is no longer simply about matching keywords. Systems can assess meaning and context. For marketers, that makes it more important to clearly establish what you know, what you have done, and where that expertise sits.

Crawler permissions. This is easily overlooked. AI companies can use different crawlers for different purposes, including training and live search retrieval. Your website's robots.txt and related settings can therefore affect whether content is available to a system.

Personalisation and context. AI answers can vary according to the query, location, search context and information available within a user's session or history. Two people asking about you may therefore receive different answers even when the sources are similar.

Hallucination. Even strong source material does not eliminate error. AI can misinterpret a source, mix up people, or generate unsupported claims. A stronger digital footprint reduces ambiguity; it does not make the system reliable.

What About Privacy?

Our professional footprint contains information we deliberately publish and information others publish about us. Employers create team pages. Conference organisers publish biographies. Publications quote us. Events are recorded. Podcasts are transcribed.

AI can bring those fragments together in ways that no individual source ever did.

Publicly available information does not always mean deliberately published for AI consumption. We face a tension: we want enough credible sources in the public domain for AI to represent us accurately, without assuming everything about us needs to remain permanently discoverable.

The answer is not to disappear. It is to be deliberate about what we make authoritative.

When the Wrong Story Gets Repeated

The more difficult problem is an inaccurate claim repeated often enough to look credible.

A former job title appears in an old biography. Another conference copies it. A third publication repeats it. Suddenly, one outdated detail looks like corroborated evidence.

If an AI system reproduces it, the instinct may be to "correct the AI." Start somewhere else.

Trace the claim back to its sources. Update what you control. Ask third parties to correct inaccuracies where possible. Then reinforce the current version through authoritative, relevant sources.

Your Footprint Is More Than What You Write

There is another shift marketers should pay attention to.

Your professional footprint is not limited to written content. A conference presentation can be recorded. A podcast can be transcribed. A panel can appear online. A webinar can be captioned. An interview can be quoted years later.

As marketers, we spend our careers thinking about content ecosystems, content repurposing, and multimedia content; this should sound familiar.

The difference is that we are now part of the ecosystem ourselves.

A Better AI Reputation Audit

I would not recommend obsessively checking what every AI system says about you. Instead, make it an occasional professional audit:

Query multiple tools. Ask two or three systems the same question about you.

Compare. Look for differences in accuracy, consistency, and currency.

Trace. If something is wrong, identify the source of the error.

Strengthen the evidence. If something important is missing, create clear, credible public evidence that establishes it.

Check the plumbing. Review your website's robots.txt and relevant AI-crawler settings. You may be limiting retrieval without realising it.

For me, the interesting shift is that the job is no longer simply to manage what people can find about us. It is to understand how different systems interpret what they find.

We have always known that reputation is contextual. The same experience can be obvious to one audience and invisible to another.

AI does not remove that problem. It scales it. And perhaps that is the question worth asking the next time an AI system tells you who you are: What did it see that I gave it, and what did it miss?

Author: Moumita Das Roy, Commercial & Communications Manager at Dulux New Zealand, 8th September 2026