brand.context: A Machine-Readable Standard for the AI Decision Stage
AIVO Evidentia · Working Paper WP-2026-04 · April 2026 DOI: 10.5281/zenodo.19588522
Abstract
This article announces the publication of brand.context — a machine-readable brand context standard designed for consumption by AI agents during commerce and purchase recommendation tasks. brand.context is the implementation layer of the AIVO Evidentia Decision-Stage Filter Taxonomy (WP-2026-01), translating eight evidence-grade filter types into a structured JSON-LD schema that brands publish at a predictable location on their own domains.
The infrastructure gap
The AIVO Evidentia Filter Taxonomy (WP-2026-01) identified eight structurally distinct mechanisms by which AI models eliminate brands at the decision stage of buying conversations, derived from 7,000+ four-turn buying sequences across 160+ brands on ChatGPT, Gemini, Perplexity, and Grok.
What the taxonomy documented is a measurement problem. What it implied is an infrastructure problem.
Brands whose content architectures do not provide structured evidence against the operating filter are eliminated regardless of brand strength, awareness, or AI visibility. But the current brand content stack — websites, JSON-LD product markup, Open Graph tags — was designed for human readers and search crawlers. None of it addresses the decision-stage layer. None of it tells an agent what a brand can demonstrate against the criteria that determine whether it gets recommended.
brand.context is the missing layer.
The standard
A brand.context file is a JSON-LD document hosted at /.well-known/brand.context on a brand's primary domain. It comprises four sections:
The Brand Identity Layer anchors the agent to the correct brand entity at T1. The Decision Filter Layer contains one structured object per Evidentia filter type, with claim and evidence pointer fields for each. The Competitive Positioning Layer encodes the brand's relative category position, including honest declaration of where it trails. Schema Metadata provides provenance, taxonomy version, and DOI linkage.
Every field maps directly to a filter type in the published taxonomy. Nothing in the schema is invented for the file — it is the diagnostic output of the AIVO methodology translated into machine-readable form.
Present benefit and forward position
The immediate value is retrieval-augmented generation performance. Retrieval-augmented systems including Perplexity and Bing-backed models actively consume structured web content. A correctly structured brand.context file at a predictable domain location receives preferential treatment over unstructured content of equivalent quality — the same structural data authority logic that governs JSON-LD for Google's knowledge graph.
The forward position is context graph participation. Platform-level signals from Perplexity, OpenAI, and Google indicate that agentic shopping infrastructure with cooperative context layers is under active development. brand.context defines the brand-side contribution to that infrastructure ahead of the systems that will consume it.
The standard is honest about this timeline. No major AI platform today crawls specifically for brand.context files. The value delivered now is real but indirect. The value when agentic infrastructure opens will be direct and immediate for brands that are already positioned.
Publication details
Working Paper WP-2026-04 is published open access on Zenodo. The full schema specification — including all filter-mapped field tables — is included in the paper. The Evidentia taxonomy it implements is published as WP-2026-01 at DOI: 10.5281/zenodo.19401584.