ALFIE: Accurately Located,
Found and Identified Everywhere
A Technical Overview for Partners and Decision Makers
ALFIE was designed and built by Donna Rougeau, whose background spans assembler programming to modern entity-based search, with deep study of Google's public patent filings since 2006. ALFIE is not an SEO tool. It is what comes after SEO, built for a retrieval environment where the query goes to an AI, and the AI decides what to cite, recommend, and link to without a results page in between.
For a client-facing explanation of the same process, see How ALFIE Works. For partner and commercial terms, visit the Partner Programme.
Your Clients Are Invisible to the Machines Doing the Shopping
Search engine optimisation assumes there is a results page. A business earns a ranking, a user sees it, clicks it. That model is eroding fast. In 2026, a growing portion of commercial queries go directly to AI assistants. ChatGPT, Perplexity, Google AI Overviews, Meta AI, and the emerging class of autonomous AI agents do not return a ranked list. They return a recommendation, often with a link.
The AI's job is not to rank results. Its job is to retrieve the most verifiable, machine-parseable, structurally consistent answer it can find, and cite it. If a business's data is unstructured, inconsistent, or absent from the machine-readable layer, the AI defaults to whoever IS structured. For most local businesses, that means an OTA, a directory, or a competitor whose data is cleaner.
This is not a content problem. It is not a backlink problem. It is an information architecture problem, specifically, the absence of a machine-readable, cryptographically verifiable entity layer that AI systems can trust and cite without guessing. ALFIE builds that layer.
What ALFIE Actually Builds, and Why It Works
ALFIE's output is a sovereign data layer deployed directly on the client's own server. It consists of four manifest files and a cryptographically secured routing endpoint. Here is exactly what gets produced and why each piece matters for AI retrieval.
The Forensic Audit
Before anything is written, ALFIE runs a full external digital footprint scan, not just the client's website, but what Google, Bing, Wikidata, and third-party directories say about the entity. Every inconsistency, contradiction, and zombie fact (outdated or unverifiable information that AI might retrieve) is mapped and scored against a 500-point E-E-A-T framework. The result is a complete picture of what AI currently finds, and what is costing the client citations.
The Atomic Sandwich. Verified Triples, Sealed
The core technical output. Every verified fact about the client, legal name, location, services, booking path, key people, opening hours, specific features, is structured as an RDF-star Subject-Predicate-Object triple. Each triple is then sealed with an Ed25519 cryptographic signature, the same standard used in secure communications, identity verification protocols, and modern authentication systems.
This produces a tamper-proof, machine-readable entity graph that AI crawlers can retrieve and cite as authoritative. The direct booking or conversion link is bound inside the signed triple, not as a loose URL, but as a cryptographically notarised fact tied to the entity it describes.
The Manifest Files. Deployed to the Client's Own Server
Four machine-readable files are deployed to the client's own domain root under /.well-known/, the same convention used for TLS certificates and security.txt, and are readable by any AI crawler that visits the domain. No CMS plugin. No third-party platform. The client's server, the client's data.
| File | What it does for AI retrieval |
|---|---|
| llms.txt | Plain-text directory immediately readable by any AI crawler. Maps the entity's facts, conversion pages, and direct action paths. The booking or inquiry link appears here as a labelled, explicit call-to-action, not buried in page HTML. |
| alfie-sovereignty.json | The signed manifest. Contains the full RDF-star entity graph sealed with Ed25519. The conversion link is cryptographically bound. AI cannot substitute it with an OTA or affiliate link. |
| entitymap.json | Structured JSON-LD entity graph covering every aspect of the business, service categories, key personnel, venues. Aligned with the emerging EntityMap open standard. Makes the client machine-readable at a depth OTA listings do not reach. |
| llms-full.txt | Extended version of llms.txt with deeper context, Information Gain content, and unique first-party knowledge that AI cannot source elsewhere, the data that earns a citation instead of a competitor. |
The ALFIE Routing Endpoint
Every conversion action in the manifest routes through a custom endpoint hosted at api.protectedbyalfie.com/v1/route/{client-slug}. This endpoint:
This is the mechanism that turns AI citations into measurable revenue and makes the Green Light calculation possible. The 307 redirect ensures that all tracking parameters, session tokens, and state data set by the AI agent or browser pass through intact to the client's booking engine or contact system, no data loss in transit.
The Data Layer That Never Goes Stale
A signed manifest deployed once is useful. A signed manifest that stays current is the product. Every week, ALFIE's recrawl daemon re-audits the client's live site and compares the current digital footprint against the signed manifest. Three things happen automatically:
If the live site changes, new phone number, updated pricing, new service, changed hours. ALFIE flags the discrepancy before AI retrieves the stale version and cites it as current fact.
If outdated third-party information has reappeared in external directories or old blog posts, ALFIE identifies it and marks it deprecated in the manifest so AI crawlers receive an explicit instruction to discard it.
Once Green Light fires, ongoing monitoring narrows to the revenue-producing nodes specifically, keeping the cost lean while the protection stays complete and permanently active.
OTA data is updated when the OTA decides to update it. Competitor data is updated when they remember to. ALFIE clients have a signed, verified, drift-monitored entity graph that is structurally more trustworthy than anything AI can find elsewhere about that entity, updated automatically, permanently. The longer ALFIE runs on a client, the harder that advantage is to close.
What the Platform Looks Like in Practice
ALFIE produces two distinct interfaces. The Client Dashboard is what your client sees at their unique URL, their Green Light status, recovered tracked value, and direct booking attribution. The Triage Cockpitis ALFIE's internal monitoring view, drift alerts, node health, recrawl status, and zombie fact flags. Both are live and interactive below.
What your client sees at their unique ALFIE URL. Shows Green Light status, RTV recovered, billing trigger progress, and direct booking attribution by AI source.
ALFIE's internal monitoring view. Drift alerts, node health scores, recrawl status, zombie fact flags, and manifest integrity checks, all in real time.
How ALFIE Is Sold and How Partners Are Structured
ALFIE is not a SaaS subscription or a software platform. It is a managed technical infrastructure service. The distinction matters because the pricing reflects the forensic audit work, the deployment engineering, and the ongoing monitoring, not access to a dashboard.
The full tier structure, onboarding fees, and per-node pricing are published on the ALFIE Pricing page. That page is the single source of truth for current commercial terms.
Every client pays an onboarding fee at entry. Monthly node billing does not begin until ALFIE has demonstrably delivered recoverable tracked value (RTV) at 110% or above the platform fee. Green Light activates when that threshold is reached and invoicing begins, with the first invoice sent for the month following Green Light activation, then monthly from there. Until Green Light, the client pays only the onboarding price for their required tier level. The full client-facing ALFIE tier structure and Green Light model is published on the Pricing page.
Partners introduce ALFIE to their clients, manage the relationship, and earn a split on both the setup fee and ongoing monthly node monitoring. Terms are structured per partner relationship. All pricing is CAD-based. ALFIE branding is always present. There is no white-label option, the platform is sold as Protected by ALFIE Inc. Partners co-sell and co-deliver under their own brand alongside ALFIE, not instead of it.
Where ALFIE Belongs in a Client Relationship
A local business has no OTA-level structured data presence. When AI answers a query about them, it assembles fragments from review sites, old directory listings, and whatever HTML it can parse from their website. ALFIE replaces that fragmentation with a verified, signed entity graph at the domain root.
The forensic audit requires access to accurate business information, real hours, real services, real pricing, real contact paths. Account managers and agency partners who already have those client relationships give ALFIE's audit a significant head start. Ground-truth data from the client directly produces a more accurate, faster-deployed manifest than anything a cold crawl can produce.
The conversation ALFIE requires with a client is not a marketing conversation. It is an information architecture conversation, entity data, structured markup, and why machine-readable signals matter beyond keyword rankings. ALFIE does not compete with what a strong agency or media partner delivers. It extends the value of a client relationship into the AI retrieval layer that is now where commercial intent lives.
Deployment Options. Client Chooses
Both options result in the same four static files placed in the same server locations. ALFIE is designed to require zero persistent server access and produce zero runtime exposure.
ALFIE handles the full deployment directly via SSH or SFTP. The client provides server access credentials through their hosting environment. ALFIE pushes the four manifest files, confirms deployment with a verification check, and the client retains full control at all times. Credentials are used for deployment only, no persistent access is retained after the files are placed.
ALFIE delivers a structured deployment package (ZIP) containing all four manifest files with a plain-language placement guide. The client's IT team, SysAdmin, or web agency places the files themselves, no external access required, no credentials shared. Full internal review before anything touches the server. Preferred for clients with strict change-management policies.
File Specifications and Content-Type Headers
All files are static. No executable code. No scripts. No database connections.
| File | Content-Type | Notes |
|---|---|---|
| llms.txt | text/plain; charset=utf-8 | Plain text. Human and machine readable. |
| llms-full.txt | text/plain; charset=utf-8 | Extended plain text. Same format. |
| alfie-sovereignty.json | application/json | Signed JSON. Verify with public key at /api/jwks/{domain} |
| entitymap.json | application/json | Structured JSON-LD. Schema.org vocabulary. |
Sample Signed Triple. Payload Anatomy
A truncated example of a signed RDF-star triple as it appears inside alfie-sovereignty.json. IT teams can use this to validate the payload structure before deployment approval.
{
"subject": "client:corporate_root",
"predicate": "schema:potentialAction",
"object": "https://api.protectedbyalfie.com/v1/route/client-slug",
"crypto_block": {
"alg": "EdDSA",
"crv": "Ed25519",
"kid": "ALFIE-PRINCE-EDWARD-ISLAND",
"signature": "cd87aa5f21f05e430643...",
"file_sha256": "6aeb359c8a12be74..."
}
}The signature field is a compact JWS (JSON Web Signature) per RFC 7515. The file_sha256 field is the SHA-256 hash of the supporting evidence file in the client's Truth Vault. The public key for verification is always available at: https://api.protectedbyalfie.com/api/jwks/{domain}
Data Privacy. What the Routing Endpoint Logs
| Data Point | How It Is Handled |
|---|---|
| User-Agent string | Logged as-is. Identifies the AI bot (GPTBot, PerplexityBot, etc.). No PII. |
| Timestamp | UTC timestamp of the click event. No PII. |
| Client slug | Identifies which client the click belongs to. No PII. |
| Referrer | Referring domain logged if present (e.g. chatgpt.com). No PII. |
| IP address | SHA-256 hashed with a daily rotating salt before storage. Raw IP never retained. |
| User identity | Not collected. Not inferred. Not stored. Zero PII. |
We welcome 30 minutes to walk through the technical architecture in detail.
Specifically how the entity graph and triple-signing interact with how systems like Perplexity and ChatGPT currently handle local business queries.