The Closed-Loop Entity Verification System: How to Build One in 90 Days

Most advice about entity building is either too abstract or too incomplete. "Create a consistent online presence" is abstract. "Add JSON-LD schema to your website" is incomplete, because JSON-LD without the external nodes it points to is a claim without confirmation.

What actually works is a closed loop. A system where each node verifies the others, where the whole is more credible than any individual part, and where every new piece of content or credential you add strengthens the entire network.

This is the system I built for my own identity and for my three companies. I'm documenting it as a 90-day protocol because that's roughly how long it takes to get all the nodes active and cross-linked.

The loop has three layers. Understanding each layer matters before you start building.

Layer 1 is your domain. Your website is the hub. Everything else is a spoke. The hub declares your identity through JSON-LD schema and rel="me" links. It is the canonical source.

Layer 2 is your verified profiles. ORCID, LinkedIn, Zenodo, OSF, Google Scholar if applicable. These are platforms with their own verification processes, meaning they add credibility your domain alone cannot manufacture.

Layer 3 is third-party corroboration. Institutional mentions, government records, media coverage, client references. These are the hardest to obtain and the most valuable, because they come from sources with no incentive to fabricate the reference.

A closed-loop entity verification system means Layers 1 and 2 are fully cross-linked, and Layer 3 is actively growing. Let me walk through how to build this in 90 days.

Days 1 to 15: Foundation

Register your ORCID. Go to orcid.org and create an account. This takes fifteen minutes. Fill in your name (use your full legal name and any known variations as aliases), your current institutional or company affiliation, and at least one work. The work doesn't have to be a peer-reviewed paper. A book, a technical report, a presentation, a documented project. Anything that represents your professional output. Make your ORCID profile public.

Why ORCID first? Because it's the most trusted individual identifier in the academic and research ecosystem, and AI systems trained on that ecosystem weight it heavily. Even if you're not an academic, ORCID is open to any professional who produces documented work. I used it to register 558 titles in my publishing catalog.

Register with Zenodo or OSF. Zenodo is CERN's open research repository. OSF is the Open Science Framework. Both allow you to deposit documents, presentations, reports, and other outputs with a DOI (Digital Object Identifier). The DOI is a persistent, machine-readable identifier that AI systems recognize as a high-trust reference.

Deposit something this week. A summary of your methodology. A technical report. A presentation deck. Whatever is already documented and can be made public. Connect your Zenodo or OSF account to your ORCID, so the platforms share a verified identifier.

Audit your domain. Before you can link outward, your domain needs to be the kind of hub worth linking to. Check three things: Does your site load over HTTPS? Does it have a clear, consistent identity declaration (your name, your company, what you do)? Does it have at least one content page that demonstrates expertise, not just a services list?

Days 16 to 30: Schema and Links

Implement JSON-LD schema on your domain. If you're a Person entity, implement @type: "Person" schema on your homepage and about page. If your company is the primary entity, implement @type: "Organization". Include the sameAs array pointing to your ORCID, LinkedIn, Zenodo, and any other verified profiles from the previous stage.

The minimum viable schema for a Person:

{
  "@context": "https://schema.org",
  "@type": "Person",
  "name": "Ibrahim Anwar",
  "alternateName": "Hibranwar",
  "url": "https://hibranwar.com",
  "jobTitle": "Practitioner & Director",
  "sameAs": [
    "https://orcid.org/YOUR-ORCID",
    "https://linkedin.com/in/YOUR-PROFILE",
    "https://zenodo.org/communities/YOUR-COMMUNITY"
  ]
}

Add rel="me" links. In your site's <head>, add <link rel="me" href="..."> for each external profile in your sameAs array. In your footer, add visible links to the same profiles with rel="me" on the anchor tags.

Update external profiles to link back. Go to each external profile and add your domain URL to the website field. LinkedIn has a "Website" field in contact info. ORCID has a "Websites & Social Links" section. Zenodo's profile allows website links. Every reciprocal link you add closes a verification loop.

Days 31 to 60: Content and Documentation

The schema and links create the structural layer. Content creates the evidence layer. Evidence is what AI systems use to characterize what you actually do.

Document your work publicly. For my company Witanabe, this means 60+ projects documented on witanabe.com/blog. Real photos. Technical descriptions. Context about the problem solved and the outcome delivered. Not marketing copy. Documentation.

For Arsindo, it means procurement records, business permits, and project documentation organized in a way that's accessible on the domain.

For my individual identity, it means essays like this one, published on a domain I own, with consistent attribution to my name and ORCID.

The target for this stage: at least 10 pieces of documented work published and publicly accessible. These become the evidence base that AI systems use to verify that you do what you claim to do.

Deposit on Zenodo or OSF. For each significant piece of documented work, consider depositing a version on Zenodo or OSF with your ORCID in the creator field. This creates machine-readable links between your output and your verified identifier. Each deposit accumulates in your citation record and strengthens the author-work relationship in the knowledge graph.

Establish Google Business Profile. If you have a physical business address, verify it on Google Business Profile. Keep the name, address, and phone number exactly consistent with what appears on your website and in your JSON-LD schema. Inconsistency here creates disambiguation problems.

Days 61 to 90: Corroboration and Amplification

The first two phases build what you control. Phase three is about getting external validation you don't control.

Identify institutional relationships worth documenting. For me, this includes EFEO Paris and KPK as institutional client relationships. These carry significant verification weight because these institutions have no commercial incentive to fabricate an association. The fact that they're documented in public records creates third-party corroboration that a machine can find independently.

If you have government clients, industry associations, academic collaborations, or media coverage, identify each one. Create a page on your domain that documents these relationships factually. Not marketing language. Documented relationships with verifiable references.

Submit to relevant directories. Industry association directories, professional body registries, government vendor lists. These are low-visibility but high-trust sources that contribute to the entity graph. The KADIN membership directory. Your industry's trade association. Any government procurement registry where your company is listed.

Trademark and formal registrations. If you have a trademark, your registration number (like my IDM001337019) is a verifiable artifact in a government database. This is one of the highest-trust corroboration signals available because it's issued by a government authority after verification.

Maintaining the Loop

A closed-loop entity verification system is not a one-time build. It requires maintenance.

Every time you establish a new external presence, add it to your sameAs array and rel="me" links. Every time you produce significant documented work, deposit it on Zenodo or OSF. Every time you have an institutional relationship, document it.

The loop gets stronger with each iteration. Each new node adds verification paths. Each new verified document adds to the evidence base. The compound effect is real and significant over 12 to 24 months.

I'm 18 months into building this for my own identity and my companies. The signals are compounding. The entity verification is becoming more solid. The AI citations, when they happen, reference a verified entity rather than a claimed one.

What This Looks Like in Practice

For someone starting from zero, 90 days gets you:

A verified ORCID with at least 3 to 5 documented works. A Zenodo or OSF account with deposited documents. A domain with JSON-LD schema and rel="me" links. At least 10 pieces of documented work on your domain. Reciprocal links from your main external profiles. Entry in at least one or two third-party directories or registries.

That's not complete. Entity building takes years, not months. But 90 days is enough to get the loop closed, to have the basic verification circuit functioning. After that, every new addition amplifies what's already there.

The businesses that start this now are building infrastructure that compounds. The ones that wait are falling further behind with each month, not because the gap in quality grows, but because the gap in verifiability grows.

Build the loop. Close it. Then keep adding to it.