Building my own person entity for Google and AI search

The goal

Make “Nishaka Avindya” resolve to one clear person entity: an SEO Specialist in Sri Lanka, working on AI search visibility. And make search engines and AI engines retrieve that description from sources I control, rather than guess.

The target is retrieval, not recall. AI engines that search the live web (Google’s AI features, ChatGPT with search, Perplexity) can be influenced by what is published and indexed today. What a model remembers from its training data can’t be, so it isn’t measured as a goal here.

The starting point

A new name with almost no search footprint: no personal website, and no structured data describing me.

The principles I worked to

The step-by-step method stays private. These are the principles behind it:

1. One identity, stated the same way everywhere. A single self-description, repeated word for word on the site, in structured data and across public profiles. Inconsistent descriptions are what make entity resolution fail.
2. A hub that everything points back to. One canonical home for the entity. Every profile links to it, and it links back to every profile.
3. Machine-readable identity. Structured data that states who the person is, what they know, where they work and where else they appear, all connected to one identifier.
4. Proof of expertise on my own domain. Original research, technical guides and working tools, published first on the hub and not on third-party platforms.
5. Measure retrieval, not assumptions. Check what search and AI engines actually return for my name on a fixed schedule, and record it.

First measurement: September 28, 2026

What I learned so far

• Retrieval works before recall. Both engines described me accurately by searching the web, not from memory. The pages are doing the work.
• The hub is the primary source. Most citations pointed to this site, not to third-party profiles. Owning the source means owning the description.
• Citations can attach to the wrong claim. Google’s answer stated my employer and university correctly, but attached a link to this site to my employer’s name and a LinkedIn link to my university. The facts were right and the links were loose. Every important fact needs a page that states it clearly, next to the entity it belongs to.
• Engines repeat your framing. ChatGPT described my positioning almost in the words this site uses. How you describe yourself is how you get described.

Results over time

This page is updated after each check. The next one is due on October 28, 2026.

What this means for your brand

The same principles apply to any business or professional that search and AI engines don’t yet describe correctly: one consistent identity, a hub that everything points to, machine-readable facts, proof on your own domain, and measurement on a schedule. That is the core of my GEO / AEO work.