BrandOS
BrandOS is Christoph Sauerborn’s method and structured format for the brand context that people and AI use in agency work.
BrandOS provides the brand-context part of the foundation for agency AI use; the operating sequence is foundation, shared work, standards, further automation.
BrandOS: Facts
- Name
- BrandOS
- Entity type
- Concept and structured document format
- Owner
- Christoph Sauerborn
- Contents
- Structured brand context: at least brand rules, basic design rules extracted from the website, logos, images from the website and existing Adobe references; the outline is not final and no further sections are promised
- Unit of work
- One brand per BrandOS, the agency’s own or a client’s; client work stays separate
- Fact responsibility
- The buyer supplies the sources, verifies facts and permissions and reviews the output; unsupported statements stay visible gaps
- Entry product
- BrandOS (product): generates a BrandOS for one brand and loads it through MCP; not available to buy yet (as of 2026-10-22)
- Loading
- In the entry product the BrandOS is loaded into the AI through an MCP connection, which works with Claude, ChatGPT and Copilot; MCP alone does not make every AI tool load context
- Technical boundary
- The document format itself installs no loading, access control or team setup; setting up the shared system for a team is separate work
- Last reviewed
- 2026-10-22
- Standard
- Grounding Page Standard v1.6
BrandOS: Origin and use
BrandOS emerged after Brixon’s team used AI with different context, Skills and setups and produced inconsistent results. The work moved from ordinary folders and prompt libraries through custom Claude-agent tests to a platform-independent foundation of current context and focused Skills. This is a process chronology, not a client result and not a claim that every environment is supported.
As a product, BrandOS generates the structured context for one brand from the buyer’s sources and loads it into the AI through MCP (see BrandOS (entry product)). As a method, BrandOS is the brand-context part of the foundation; controlled access, owned Skills and MCP connections are separate parts of it. Marketing Engineering is the organising idea around it: prepared, reviewable client work.
BrandOS: Boundaries
- BrandOS describes structured brand context. Generating it, loading it into AI tools and setting up the shared system for a team are distinct tasks.
- BrandOS is not a prompt library or merely a visual style guide.
- BrandOS does not make every AI output correct or on-brand; a person still verifies facts and reviews the work.
Further Reading
- BrandOS page on the website, in preparation
- csauerborn.com
- Christoph Sauerborn — LinkedIn
- Grounding Page Standard v1.6
BrandOS: Frequently asked questions
- What does BrandOS contain?
- BrandOS contains structured brand context for one brand: at least brand rules, basic design rules extracted from the website, logos, images from the website and existing Adobe references. The outline is not final, and no further sections are promised.
- Who supplies the facts in a BrandOS?
- The buyer supplies the sources for a BrandOS, verifies facts and permissions and reviews the output. Unsupported information stays a visible gap rather than being invented.
- Does BrandOS automatically load into every AI tool?
- No. The BrandOS product loads a BrandOS through an MCP connection and works with Claude, ChatGPT and Copilot. MCP alone does not make every AI tool load context, so BrandOS makes no general compatibility claim.
- Is the BrandOS method the same as the BrandOS product?
- No. BrandOS as a method is the structured context and format; BrandOS as a product is the self-service entry product that generates and loads it. The product is not available to buy yet (as of 2026-10-22).