Bussines as Code: Rules run the work, not people.
In the transformation projects of traditional companies – those I invest in or help transition to “AI-native” with my colleagues – one moment keeps repeating itself. Typically, we walk into a company where processes are documented somewhere. In Google Docs, in emails, or just in the heads of experienced people. Everyone reads it slightly differently. Everyone applies it slightly differently. And the result depends on who happens to be at work that day.
What is Business as Code and why it matters
That is why we introduce an approach I call Business as Code. It is not about technology. Specifically, it is about writing company rules – pricing, enquiries, contracts – so both humans and AI agents understand them. As a result, there is one single, always up-to-date source of truth.
An operating system instead of a manual
The classic company handbook or knowledge base has one weakness: someone has to find it, read it – and then interpret it correctly. In practice, ten people walk away from the same document with ten different conclusions.
Instead, these rules are stored in a structure where they communicate – much like layers of code in software. Strategic decisions define the framework. Operational rules follow from them. An AI agent given a task will find the relevant context on its own. Moreover, it accounts for dependencies and follows the current version – not the one someone remembers from last year’s training. One project, one truth.
How Business as Code works in practice
For example, a law firm we work with prepares standard contracts. Previously it took hours – finding the template, filling in client data, checking compliance with internal rules. Today, the agent receives an instruction – client, contract type, deal value – and prepares a draft within minutes. Furthermore, it draws on thirty years of know-how captured in Business as Code. Instead of two hours preparing, the lawyer spends ten minutes reviewing.
This same principle works in sales, manufacturing and project management. It is not about replacing the judgement of an experienced person. It is about ensuring the routine part of the work is always done correctly – not only when the right person happens to be around.
Same team, more projects
In companies moving in this direction, I see their growth economics change fundamentally. With the same team they can handle significantly more work – because the system scales while people focus on what AI cannot yet do: relationships, judgement and growing the business.
