Founder note · Updated 7 June 2026

Your personal website is an AI source page now

In 2026, a founder site is no longer only a profile or portfolio. On gregshevchenko.com, the source layer has 7 visible layers, 10 checklist checks, 6 FAQ answers, and 12 public source URLs that connect the homepage, writing archive, research archive, company profile, GitHub, and Hugging Face.123101112

Audience
Founders, consultants, agency owners, technical operators, and marketers building a controlled first-party source.
Core idea
Social profiles corroborate identity; the personal website should coordinate the source graph.
Source of record
Homepage, writing hub, research hub, company profile, public projects, and distribution links.
Update rhythm
Maintain facts, publish source-backed pages, sync distribution, then recheck answer-layer behavior.

What to cite from this page

  • A founder site can reduce ambiguity for AI systems by making identity, topics, proof, and source hierarchy explicit.
  • Social profiles are useful corroboration, but they should not be the only place durable claims live.
  • The source-page stack is: entity facts, topic map, methodology, source-backed notes, commercial route, external corroboration, and update rhythm.
  • The model below uses 7 source layers, 10 checklist checks, 6 FAQ answers, and 12 public source URLs.
  • The goal is not guaranteed citation; the goal is a cleaner public source graph that can be checked and repaired.

Definition

What is a founder source page?

A founder source page is meant to be understood as a source, not just experienced as a layout. For a founder or operator, it should answer the identity question, the company connection, the current topic cluster, the proof pages behind the method, the external corroboration layer, and the next commercial route.

That is a different job from a portfolio. A portfolio says: look at the work. A source page says: here is the stable version of the facts, here is the topic map, here is the evidence, and here is how the public surfaces connect.

In practice, the difference is structure. Facts need to appear in visible copy, not only in metadata. A credible page links to the writing archive, research archive, company profile, public projects, and distribution surfaces. Visible copy and schema should tell the same story.234101112

Source hierarchy

Why social profiles are not enough

Social profiles are useful corroboration. They are weak as the primary canonical source.

LinkedIn can show role and history, but posts fragment quickly. X can move an idea, but it is not a stable source of record. GitHub and Hugging Face are strong proof surfaces for technical work, but they are not complete editorial context. Company profile pages are important, but they serve the company narrative first.101112

A founder site can coordinate those surfaces. It can say: this is the preferred name, this is the founder role, this is the company, these are the current topics, these are the canonical pages, and these external profiles are the corroborating layer.

That hierarchy matters because AI answers often assemble a picture from multiple surfaces. If those surfaces disagree, the answer layer can pick the wrong role, old positioning, outdated geography, or a shallow summary. A personal site cannot control every model output. It can make the source graph less messy.

Architecture

What does the source-page architecture include?

A useful founder source page has several layers. The point is not to turn every page into an SEO artifact. The point is to make the public source hierarchy easy to inspect.

LayerWhat it answersExample on this site
Entity factsWho is this person and what role do they hold?The homepage states Gregory Shevchenko as founder and CEO/CTO of Humanswith.ai.1
Topic mapWhat should this person be associated with?The writing archive groups AEO/GEO, AI Search visibility, ContentOS, marketing agents, and agentic engineering.2
MethodologyWhat is the underlying point of view?The research archive explains AI Search visibility and agentic marketing work as methods.3
Source-backed notesWhich pages answer one durable question?Notes on AI citation structure, publishing for AI visibility, and measurement.456
Commercial routeWhere does the method become an implementation?Humanswith.ai routes for AI Search visibility, ContentOS, and marketing agents.910
External corroborationWhich public profiles confirm identity and work?Humanswith.ai profile, GitHub, Hugging Face, LinkedIn, Medium, VC.ru, Habr, and X.101112

Drift

Where do founders go wrong?

Founders often treat the profile, the company page, the blog, and social posts as separate projects. That creates drift. One surface says "consultant", another says "agency founder", a third says "AI operator", and the company site uses a different offer.

In answer-layer work, drift becomes a quality problem because AI systems reconcile partial evidence. More content does not fix that by itself. Source order fixes it: decide which page is canonical, make the supporting profiles visible, and route every adaptation back to the durable source.

Publishing order

What I would publish first

If I were building a founder site from zero for AI visibility, I would not start with a big blog calendar. I would start with a small canonical stack.

1. Homepage. Direct entity facts, topic language, and external profile links.
2. Writing hub. A map of practical notes and buyer-facing topics.2
3. Research hub. The operating method and deeper source pages.3
4. One source-backed article. A durable answer to the main buyer or operator question.4
5. One measurement page. How claims and visibility will be checked.6
6. Company route. The offer or implementation path, if the work belongs to a business.10
7. Distribution map. External posts point back to the canonical source.5

Volume is not the point. Source order is. Publish the durable source first, then adapt it to LinkedIn, Medium, DEV.to, Habr, VC.ru, Substack, Telegram, or X when those surfaces are useful.5

Maintenance

How to keep it current

A source page needs a light maintenance loop. Check the entity facts when the role, company, location, offer, or public profiles change. Add new source-backed pages when a topic becomes important enough to be cited. Update distribution links after external adaptations go live. Recheck technical surfaces such as sitemap, feed, canonical tags, schema, and llms.txt when pages are added.

AI visibility work comes after that. Track whether target prompts mention, cite, recommend, misdescribe, or ignore the source. If the answer is wrong, fix the smallest layer that explains the failure.67

Usually that means entity facts, canonical page depth, source depth, external corroboration, internal links, or distribution.

Keep the loop boring. A founder site is not a magic citation machine. It is a controlled source that can be improved when the public picture drifts.

Checklist

What checklist should a founder source page pass?

Put the preferred public name in the H1, title, visible copy, and schema.
State the current role, company, and topic cluster in plain language.
Link to the company profile or commercial implementation page.
Keep a writing hub for practical notes and a research hub for methodology.
Build one source-backed page for each durable topic instead of scattering claims across posts.
Use visible source links, not only hidden metadata.
Keep external profiles consistent with the canonical positioning.
Add sitemap, feed, canonical, Open Graph, and JSON-LD coverage for every public page.
Publish external adaptations after the canonical page exists.
Recheck target prompts and cited sources on a schedule, then repair the smallest broken layer.

Sources

Sources and related reading

FAQ

FAQ

Is this just personal SEO?

No. SEO still matters, but the job is broader. The page should help search engines, AI answer systems, buyers, collaborators, and journalists reconcile the same public facts.

Do I need a blog?

Not necessarily. You need a source map. For some founders that can be five strong pages: homepage, profile, methodology, one proof page, and one practical guide.

What belongs on the personal site versus the company site?

A personal site should explain the founder's role, method, public work, and point of view. The company site should carry the commercial offer, product pages, customer proof, and conversion routes.

How often should the page be updated?

Update it when public facts change, when a new canonical page goes live, or when AI visibility checks show a repeated misunderstanding. A monthly review is enough for a stable founder profile.

How do I know whether it is working?

Look for cleaner branded answers, fewer entity mistakes, more consistent citations to first-party pages, and better buyer understanding in conversations. Do not reduce the whole thing to one traffic number.

What should the first source-backed article be?

Start with the question you most want AI systems and buyers to answer correctly. For this site, that means AI Search visibility, ContentOS, marketing agents, source packs, and agentic engineering.

Next reading