Your organisation's AI knowledge, in a format that travels.
Most organisations deploying AI are building on sand. Is yours?
The problem
AI tools are increasingly capable. The knowledge that makes them useful to your organisation specifically, your standards, your approach, the context that reflects how you actually work, is a different matter.
Right now, that knowledge lives inside a vendor's platform. It was configured there, it runs there, and if you ever want to change provider, run an audit, or simply understand what your AI was told to do on a given date, the honest answer is usually that you cannot find out easily.
The problem is one of design. Knowledge and execution have been bundled together when they should be separate. Score separates them.
What Score does
Score gives your organisation a way to encode what your AI should know and how it should behave, in a format you own and control.
That knowledge lives in files that belong to you, sit in systems you already manage, and travel with you regardless of which AI tools you use. When your requirements change, you update your knowledge. When a better model becomes available, you switch without rebuilding from scratch. When a regulator asks what your AI was permitted to say six months ago, you have a verifiable answer rather than an educated guess.
Score is an open standard, published under the MIT licence. It is free to use, free to build on, and owned by nobody.
The principle
Score uses the language of music as its organising metaphor.
Orchestras are interchangeable. The Score remains.
Who it is for
Score is relevant to any organisation that is deploying AI and wants to retain control of the knowledge that drives it.
In regulated industries — financial services, legal, healthcare, and professional services — the governance and audit questions carry real consequences. Score provides the structure to answer them. For organisations outside regulated sectors, the case is simpler: the knowledge your AI draws on is a strategic asset, and strategic assets should not live inside someone else's system.
Two paths
For senior leaders and enterprise teams
Score is created and maintained by MultipleWorks, a boutique consultancy based in Hong Kong. If you are thinking through what Score means for your organisation's AI strategy, we are available for advisory conversations.
Talk to MultipleWorks →For developers
The full specification, schema, reference implementation, and examples are on GitHub. MIT licence.
View the spec →