BIOS: 33 Specs From Real Data, Not Templates

The Template Problem

Every branding agency sells the same deliverable: a PDF with your brand colors, fonts, tone of voice, and a few persona cards. It takes 6 weeks. It costs $15,000. And the moment you hand it to an AI, the AI ignores 90% of it because PDFs aren't machine-readable constraint systems.

The Brand Intelligence Operating System (BIOS) solves this. It's Step 4 of the Context-First methodology — 33 distinct specifications generated from the Data Warehouse (Step 3), structured as machine-readable constraints that AI agents consume natively.

Not 33 documents someone wrote in Google Docs. 33 specifications extracted from evidence.

The 6-Tier Taxonomy

The BIOS organizes brand intelligence into a strict hierarchy. Each tier builds on the one below it:

Tier 1: Brand Foundation — Who You Are

The anchor. The things that never change.

• Brand Ethos: The irreducible identity statement. Not a mission statement — a constraint. Celtic Knot's ethos: "Heritage connection through craftsmanship." Every output that doesn't serve heritage connection violates the BIOS.
• Brand Archetype: The psychological framework. Celtic Knot maps to Creator-Sage — craft mastery plus wisdom tradition.
• Voice & Tone: Not "professional and friendly" (every brand says that). Specific modulation rules: "Warm but never casual. Heritage-proud but never nostalgic. Craftsmanship-focused but never pretentious."
• Governance (Restraint Doctrine): What the brand explicitly refuses to do. No hard-sell language. No discount-led messaging below positioning threshold. No cultural appropriation of Irish heritage. These restraints are as important as the permissions.

Tier 2: Brand Context — How You Show Up

The visual and competitive environment.

• Visual Identity Protocol: Not just hex codes — rendering rules. When to use the heritage serif versus the modern sans. When the Celtic knot motif appears and when it would be overuse. Texture preferences. Photography direction: lifestyle over product-only, natural light, heritage settings.
• Competitive Landscape: Who you're positioned against, where you differentiate, and what competitive claims you can factually make.
• Reference Library: The canonical documents, campaigns, and artifacts that define "on-brand."

Tier 3: Customer Intelligence — Who You Serve

Deep audience understanding, derived from data.

• Customer Archetypes: Not 4 generic personas. Data-derived segments with statistical backing: purchase patterns, review language clusters, email engagement profiles, and demographic distributions. Celtic Knot's primary archetype emerged from sentiment analysis of 907 reviews — not a workshop brainstorm.
• Psychographic Profiles: Emotional triggers, values hierarchies, and decision-making patterns by segment.
• Journey Maps: Actual customer journeys from first touch to repeat purchase, with conversion friction points identified from analytics data.

Tier 4: Product & Pricing — What You Sell

Operational viability and value framing.

• Product Intelligence: Full catalog with attribute taxonomy, margin profiles, inventory velocity, and cross-sell affinity scores.
• Pricing Strategy: Not "premium positioning." Specific price-point psychology by category, discount rules (when, how much, to whom), and competitive price anchoring.
• Alternative-Cost Anchoring: How to frame value against alternatives. Celtic Knot cashmere vs. fast fashion: "You'll own this for 20 years. That's $6 per year for something made in Ireland."

Tier 5: Content & Messaging — How You Speak

Communication across every channel.

• Content Pillars: 4-6 thematic categories with weight distribution (e.g., Heritage 35%, Craftsmanship 25%, Irish Culture 20%, Customer Stories 15%, Behind the Scenes 5%).
• Messaging Strategy: Channel-specific voice modulation. Email is warmer and longer. Social is concise and visual. Ads are benefit-led with heritage anchoring.
• Campaign Archive: Every campaign that's been run, its performance, and what worked. This prevents agents from reinventing what's already been tested.

Tier 6: Operational Intelligence — How You Run

Execution rhythm and feedback loops.

• Brand KPIs: Not vanity metrics. The 5-7 numbers that actually predict business health: ROAS, repeat purchase rate, email revenue per subscriber, customer acquisition cost by channel, average order value.
• Seasonal Playbooks: Month-by-month operational cadence: when to launch collections, when to run sales (and when NOT to), holiday messaging calendars.
• Data Warehouse: The living connection back to Step 3. This specification defines how the warehouse feeds the BIOS and how the BIOS evolves with new data.

From Warehouse to Specs

The generation process is semi-automated:

1. Data Extraction: Scripts query the warehouse for each spec's required data points
2. Initial Generation: An AI agent synthesizes the raw data into a draft specification
3. Cross-Platform Review: The draft is independently reviewed by Claude, ChatGPT, and Gemini
4. Human Review: I read every spec and validate against my knowledge of the brand
5. Finalization: Approved specs are formatted as structured markdown with frontmatter metadata

Each spec follows a consistent format:

---
spec_id: BIOS-03-01
tier: 3
category: customer-intelligence
title: Primary Customer Archetype
version: 1.3.0
last_updated: 2026-02-28
data_sources: [shopify_orders, klaviyo_profiles, review_sentiment]
confidence: 0.87
---

The confidence score is critical. It tells agents how much to trust each spec. A confidence of 0.87 means "strong statistical backing, minor gaps in data." A confidence of 0.62 means "directional but not proven — use with caution."

Why 33 and Not 10

Simpler frameworks exist. "Brand purpose, audience, voice, visuals" — four specs and you're done. So why 33?

Because four specs produce generic AI output. Thirty-three specs produce output that sounds like it came from someone who's worked at the company for five years.

The difference between a 4-spec system and a 33-spec system is the difference between "We make beautiful Irish jewelry" and "Your Claddagh ring connects to a 400-year tradition from the fishing village of Claddagh in Galway, where Richard Joyce designed the first ring as a symbol of love, loyalty, and friendship after returning from years of captivity."

The deep specs — restraint doctrine, psychographic profiles, alternative-cost anchoring, seasonal playbooks — are what make an agent genuinely useful instead of generically competent.

BIOS vs. AXIS

The same 6-tier framework has a scientific variant called AXIS (Agentic eXperimental Investigation System), used for research projects like Genesis-Witness.

Same structure. Different vocabulary:

This duality is what makes Context-First methodology domain-agnostic. The same framework that sells jewelry also produces falsifiable scientific hypotheses published at CERN's Zenodo repository. The constraint architecture works regardless of domain — because intelligence is structured the same way whether you're selling cashmere or testing quantum consciousness theories.

Versioning

The BIOS isn't static. Celtic Knot's BIOS has evolved from v1.0 to v1.3.0. Infinite Awakening's BIOS was updated just last week.

Each version captures:

• What changed and why
• Which data triggered the update
• Impact on existing agent behavior
• Backward compatibility notes

When the Q4 2025 campaign data showed that video ads outperformed static by 3.2x, that insight didn't live in a Slack message. It was incorporated into BIOS Tier 5 (Content Strategy) as a channel allocation constraint. Every agent born after that update knows: default to video for Meta, reserve static for retargeting.

What This Means

The BIOS is the most commercially valuable asset in the Context-First system. It's what clients buy. It's what agents consume. It's what makes the difference between "AI-generated content" and "content that sounds like it came from inside the brand."

But it can't exist without the Data Warehouse (Step 3), it can't be consumed without agents (Step 5), and it can't improve without feedback loops (Step 12).

Every step depends on the one before it. Skip one, and the system degrades. Honor the chain, and you get 620% ROI improvement, 78% ad spend reduction, and agents that produce 85-90% first-pass completion rates.

That's not a prompt engineering trick. That's a constraint architecture operating on real data.