.brand

The brand
becomes
the agent.

A .brand file is a living MCP agent. It embodies identity, adapts to every person, learns from every interaction.

12 personas4 languages22 scenarios
The shift

Brand guidelines can't talk to AI agents.

AI agents are becoming the primary interface between consumers and brands. A 200-page PDF doesn't negotiate with a shopping assistant. A Frontify library doesn't adapt its tone for a 22-year-old in Lausanne versus a 55-year-old in Bern. The tools we have manage brand assets. What we need is something that embodies brand intelligence.

Today: static guidelines

PDFs, digital asset platforms, compliance checklists. One-directional. The brand transmits, the consumer receives. AI agents scrape and approximate. No understanding, no adaptation, no conversation.

.brand: living identity agent

An MCP server that any AI system can connect to. It doesn't describe a brand — it is one. It speaks, validates, generates, refuses, adapts, and learns. Not just colors and fonts — textures, sounds, feelings, stories, culture.

« Tu poc la marque — elle te poc back. »
Nicolas Paupe, Brigade Studio — You poke the brand, it pokes you back.
.brand agent
01

Same brand.
Every person sees
something different.

02

What the agent decided.

03

Seven layers.
Three freedoms.

An operating system for brands that think. Each layer has a defined freedom level.

01Core IntelligenceLocked
02Expression EngineControlled
03Context AdaptationControlled
04Interaction LayerAutonomous
05System NetworkControlled
06GovernanceLocked
07Value CreationAutonomous
Locked Controlled Autonomous

Six typologies exist.
None of them understand.

Building on Irene van Nes. Container, Transformative, Generative, Reactive, Participative, Evolutive. And then a seventh.

01

Container

Fixed shape, variable content. MTV, AOL.

02

Transformative

The logo itself mutates. Google Doodles.

03

Generative

Algorithmically created. Nordkyn.

04

Reactive

Responds to real-time stimuli. MIT Media Lab.

05

Participative

Co-created with users. RISD.

06

Evolutive

Changes over time. Seed Media.

07

Conscious Identity

Not just dynamic. Autonomous. A brand that perceives, decides, and learns.

Strategic architecture — Nicolas Paupe

Seven layers of a living brand.

An operating system for brands that think. Each layer has a defined freedom level — from locked core identity to autonomous real-time expression. Not rigidity. A spectrum of creative freedom with hard edges.

01Core IntelligenceLocked
Vision, values, positioning, brand DNA. The immutable nucleus. Everything downstream inherits from here. AI cannot override — only interpret within these boundaries.
02Expression EngineControlled
Tone, visual language, narrative formats, sensory palette. AI operates within a defined segment of the brand universe — generating variations that flex without breaking recognition.
03Context AdaptationControlled
Consumer typologies, cultural narratives, temporal context. The brand reads its audience. Positioning is triple: anthropological, sensory, and neuro. It doesn't manipulate — it resonates.
04Interaction LayerAutonomous
Real-time outputs — the consumer-facing interface. Content, layout, language, emotional texture adapt per individual. Every expression unique, every expression on-brand.
05System NetworkControlled
Agent-to-agent dialogue, output negotiation, cross-touchpoint synchronization. An ecosystem of agents that share context, negotiate outputs, and maintain coherence across channels.
06GovernanceLocked
Human oversight, creative direction, strategic arbitration, ethics. Humans are the custodians of meaning. Data governance: relevance, permission, weight, temporality, risk filtering.
07Value CreationAutonomous
Products, experiences, services, brand extensions. The .brand doesn't just communicate — it produces. Every interaction generates signal. A/B testing, optimization, learning loops feed back into the system.
Locked — AI cannot override. Human-defined, machine-enforced. Controlled — Adapts within defined parameters. Autonomous — Free to interpret, anchored by layers above.
Competitive landscape

Nobody is building what we're building.

The market offers platforms. We're proposing a protocol.

Frontify, Bynder, Brandpad

Digital asset management

Filing cabinets. They store brand assets, they don't understand them. They can't talk to AI agents.

brand.ai

AI brand compliance

Closest to our vision. Machine-readable guidelines + AI generation. But centralized SaaS, not a protocol. Focus is compliance, not emotional resonance.

Firsthand ($26M Series A)

Brand agents for ads

Consumer-facing, right direction. But ad-tech DNA. Their agents are campaign tools, not embodiments of brand identity.

.brand

Open protocol — identity-first, consumer-facing

An MCP standard the brand owns and deploys. Starts from brand DNA and emotional depth. Talks directly to people. Gets smarter over time.

Roadmap

What comes next.

From proof of concept to ecosystem.

Phase 1 — Now

Brand knowledge API. Structured data, semantic search, natural language queries. SBB proof of concept live on Cloudflare edge.

Phase 2

MCP brand agent. The API becomes a proper .brand file any AI system can connect to.

Phase 3

Optimization loop. Every interaction generates signal. A/B testing, performance monitoring. The brand gets smarter.

Phase 4

Multi-channel deployment. Digital, physical, spatial. Total coherence.

Phase 5

The .brand ecosystem. File format spec, marketplace, inter-brand communication.