On this page
Why an automation studio builds Knowledge Panels
A Google Knowledge Panel is the box Google shows beside the results when it is confident it knows who a person is. It appears when the pages about that person agree with each other. Our service builds that agreement on purpose: one source record, code that describes it, every profile synced to it, and outside pages that confirm it. Google decides whether to show the panel. We build what it decides from.
Most panel work is sold as a campaign: a burst of press, some profiles, a wait. We see the same problem we see in a messy CRM. Six copies of the truth, each slightly different, and a reader (here Google, or ChatGPT) forced to guess which one is right. When systems guess, they guess badly, and that is the same failure we describe in why AI gives confident wrong answers.
So we run it as a build. There is a source of truth, the downstream copies read from it, and there is a monitor. The difference shows up a year later, when a new job title or a new company goes in once and reaches every profile, instead of leaving three old versions behind for Google to trip on.
The four-layer record
Every engagement builds the same four layers, in this order, because each one depends on the layer above it.
One page on a domain you own that states who you are in plain text. The identity sentence lives here first.
Structured code on that page that tells Google, in its own format, that this is a person, what they do and which profiles are theirs.
Every profile you control rewritten to match the source: same name string, same photo, same opening words.
Independent pages that say the same thing in their own words. This is the layer Google weighs most, and the slowest to build.
The same record feeds the AI answers
When a prospect asks ChatGPT, Claude, Perplexity or Gemini who you are, the assistants that search the web read the same public pages Google reads. If those pages disagree, the answer comes back vague, merged with a namesake, or wrong. If they agree, both Google and the assistants tend to describe you the same way. One clean record serves both readers, which is why we never quote for them separately.
The evidence points the same direction. Ahrefs looked at 75,000 brands and found that how often a brand is mentioned across the web tracks its visibility in Google's AI Overviews far more closely than its backlinks do. Mentions are the confirmation layer. They are slow to earn and nobody can fake them at scale, which is exactly why they count.
How AI assistants decide whom to recommend, and why being described consistently matters more than being described loudly.
If you want the engineering view of how an assistant picks its sources, our guide on citations and grounding covers it from the builder's side. The short version for your name: an assistant can only cite what it can find and trust, and it trusts what agrees with itself.
See the description layer for yourself
This is the code layer, in a working tool. Fill in the fields and it writes the block Google reads on your own page. It is the part of the record you fully control, and it is also the part people most often get wrong, usually by claiming facts the visible page does not say.
Nothing leaves your browser. Paste the result into the page about you, and make sure every value also appears on that page as normal text.
It describes the record. Google compares it with what the page says and what other sites say. A page that claims more in code than in text is a page Google learns to ignore.
Scope, as a build plan
| Workstream | What ships | How we check it |
|---|---|---|
| Read | A written diagnosis: what Google shows signed out, what three AI assistants say, which route your name is on | Saved screenshots and dated answers, kept as the baseline |
| Source | A page about you on your own domain, identity sentence first | Plain-text facts match the code, line for line |
| Description | Person and profile-page code, with every profile you own listed | Google's Rich Results Test and a manual read of the output |
| Sync | Every profile you control rewritten to the source | A drift report, field by field, before and after |
| Confirmation | A mentions programme: interviews, podcasts, guest pieces, listings that fit your field | Each piece logged with its URL and the facts it states |
| Handover | The claim packet for your own Google account, plus the monitor | You run the claim yourself. Your account, your hands |
Your name searched signed out in more than one region, three AI assistants asked the same question, all of it saved with dates.
The page about you goes live with its code, and every fact on it is checked against what the page says in text.
Profiles rewritten one by one, with a drift report at the end showing zero differences on the fields that matter.
Outside pages that name you correctly, placed where your field actually reads.
A monthly re-run of the baseline, so you can see the record taking hold instead of guessing.
An older one on using ChatGPT to find outside pages worth earning. The method still holds for the confirmation layer, as long as every placement is real.
Proof you can check in one click
Founders and operators from software, AI and digital whose panels our team built. Click any name to see what Google shows right now. Not every panel is at the same stage, and the table says so plainly.
| Who | Where | What Google shows today | Check it |
|---|---|---|---|
| Akshay Makadiya Founder, RankLane | Rajkot, India | Panel with a description and facts | Open on Google ↗ |
| Umesh Uttamchandani Co-founder, DevX | Ahmedabad, India | Panel with profiles tied together | Open on Google ↗ |
| Harry Rao Founder and CEO, TestGrid | Atlanta, United States | Name recognised, panel forming | Open on Google ↗ |
| Vishal Virani Co-founder and CEO, Rocket | Surat, India | Name recognised, panel forming | Open on Google ↗ |
| Ilan Nass Founder, Taktical Digital | New York, United States | Panel with profiles tied together | Open on Google ↗ |
Five of the 23 panels on our record, checked against live Google on 10 September 2026.
Two clocks: ours and Google's
We will not give you a panel date, because nobody honestly can. What we can give you is a finished record inside a quarter, a monthly baseline that shows whether Google and the assistants are reading it, and the experience of 23 panels that says a strong name tends to see one within 6 to 12 months of starting.
Who it fits
There is no price on this page because the scope depends on your name, how common it is and what already exists. You get it in writing after the read. Prefer to build it yourself? The DIY build is the same four layers, with the code.