Most conversations about AI in the businesses we evaluate start with a specific tool: a chatbot for customer support, a copilot for the sales team, an automation for invoice processing. Those are reasonable starting points, but they treat AI as a feature to be added rather than a capability that changes how the whole business operates. The distinction matters because the value compounds very differently depending on which mental model a management team is using.
A feature improves a task. An operating system changes a workflow.
A chatbot that answers customer questions faster is a feature — useful, measurable, and bounded. Redesigning the support workflow so that the same underlying capability triages tickets, drafts responses, flags product issues for engineering and updates account records automatically is a different kind of change: it removes work from the process rather than speeding up one step inside it. The businesses seeing the largest returns from AI are the ones willing to redesign the workflow around the capability, not just insert the capability into the existing workflow.
The bottleneck is rarely the model
By the time a mid-market business is evaluating AI adoption, the underlying models are usually capable enough for the task at hand. The actual bottleneck is almost always data quality, process clarity and decision rights — a model can only automate a decision that is documented well enough to automate. This is why our approach treats AI adoption as a byproduct of the same foundational work we do in finance and operations: clean data and clear processes are prerequisites, not a separate technology initiative.
AI does not fix a disorganised business. It reveals exactly how disorganised it is, at a speed that makes the problem impossible to ignore.
Governance has to move as fast as adoption
Giving a system more autonomy over customer communication, pricing or financial reporting without a corresponding increase in oversight is how AI adoption turns into a liability rather than an advantage. Every AI capability we help a portfolio business deploy comes with a defined decision boundary — what it can do autonomously, what it escalates, and who reviews the outcomes. That discipline is what makes faster adoption safe rather than reckless.
Where this fits in how we invest
Technology is one of the five levers in our value creation playbook, alongside finance, operations, governance and leadership, and AI is increasingly the fastest-moving part of that lever. We look for management teams who see it the same way: not a project with a start and end date, but a capability that gets woven progressively into how the business runs finance, serves customers and makes decisions.