Context Is The Moat
Every few years, a new tool promises to unlock your data.
Self-service BI. Natural language queries. AI assistants that let anyone “ask the database anything.” The demos are always impressive. A CEO types a question in plain English and gets a chart. Magic.
Then you deploy it. The answers are subtly wrong. The chart shows a number no one recognises. Someone asks “what does ‘active customer’ mean here?” and nobody knows — including the software.
The problem isn’t the tool. It’s the context it doesn’t have.
“Revenue” can mean five things in your business. “Customer” depends on whether you count trials, whether you count seats or accounts, whether you include the enterprise deal that signed in December but started in January. “Churn” might be logo churn or revenue churn or something the previous CFO made up and nobody questioned.
These definitions are the product of years of decisions, edge cases, and political negotiations. They live in spreadsheets, in people’s heads, in the footnotes of board decks. No tool can infer them. They have to be taught.
This matters more as you shift work from people to machines.
When a human runs the report, they carry context with them. They know that one customer is excluded from the churn calculation because of a contractual quirk. They know that “revenue” in the sales dashboard means something different than “revenue” in the finance pack. They adjust, interpret, footnote.
Machines don’t. They take the data at face value and proceed confidently. The context that lived in someone’s head has to be made explicit — written down, structured, machine-readable — or the output will be wrong in ways that are hard to spot and expensive to fix.
Building this meaning layer doesn’t show up in vendor demos. It’s unglamorous, slow, and requires decisions that people have been avoiding for years.
But here’s the thing: every company has access to the same software. The same BI platforms, the same AI models, the same query engines. The tooling is commodity.
What isn’t commodity is your context: your specific definitions, your business logic, your edge cases, your institutional knowledge of why that one customer is excluded. The software accelerates. The context decides whether that acceleration goes somewhere useful.
The organisations that get value from these tools aren’t the ones with the biggest budgets or the fanciest vendors. They’re the ones who did the unglamorous work first: defining their terms, documenting their logic, creating the layer that makes any tool useful.
If you’re about to deploy a new analytics platform or AI assistant, ask: have we defined our terms clearly enough that someone — or something — outside the business could use them correctly?
If not, start there.
Related: Definitions Are Infrastructure · From Data to Information · When Someone Leaves
Connects to Library: Tacit Knowledge