Where AI adds real value
We use AI extensively in fund management and see clear practical value in the right applications. The real question is not whether it matters, but where it genuinely improves judgement, execution and capacity.
AI is already useful in many businesses. We use it extensively in fund management, where it materially improves speed, synthesis and the quality of first-pass analysis. That has reinforced a broader view: the most valuable applications are rarely the loudest or the most ambitious. They are the ones that solve a real operational problem, precisely and consistently.
AI is often discussed too broadly. Some treat it as a universal answer. Others dismiss it because the hype has outrun the reality. Both positions miss the point. Like most tools, AI is valuable when applied precisely to the right problem. The businesses that benefit most are not the ones with the grandest AI narrative. They are the ones that understand where time is being lost, where capable people are carrying unnecessary manual load, and where better access to information would improve decisions.
Start with the bottleneck, not the technology
The biggest mistake with AI is to start with the technology rather than the problem. A business hears that AI can automate workflows or improve productivity, then starts looking for places to use it. That usually leads to low-value experimentation and a lot of activity without much practical gain.
A better approach is to start with friction. Where are capable people spending time on work that is repetitive, manual or structurally low-leverage. Where is important information buried in documents, spread across systems or too slow to surface. Where are teams relying on inconsistent judgement because the underlying data is hard to access, compare or interpret. Those are usually the right starting points.
In practice, the early wins tend to be specific. Faster synthesis of large volumes of information. Better document handling. Smarter internal search. More consistent first-pass analysis. Better drafting support. Faster triage. Pattern recognition across fragmented data. None of that sounds dramatic in isolation. But if it sharpens judgement and frees capable people to spend more time on commercial thinking, customer conversations and operating decisions, the cumulative value is meaningful.
Used well, it creates real capacity
One of the more practical benefits of AI is that it creates room. In many growing businesses, the constraint is not lack of demand. It is that strong people are stretched too thin and management bandwidth is consumed by work that should be faster or easier.
In that context, AI can be genuinely useful. It helps teams handle more complexity without losing responsiveness. It makes commercial teams quicker to prepare and follow up. It improves management visibility. It reduces the manual effort required to produce useful operating information. It helps founders and managers spend more time on decisions that actually require their judgement.
That is the healthier frame. A business that uses AI well should become sharper, faster and more scalable. The point is not to deploy technology for its own sake. It is to build a stronger operating model with more capacity in the right places.
Judgement still sits with people
AI can support judgement. It does not replace it. This is where precision matters most. In many operating environments, particularly lower mid-market businesses, the decisions that matter are contextual. They involve trade-offs, timing, customer nuance, team credibility and a practical understanding of how the business actually works. Those are not decisions to delegate casually.
That does not reduce the value of AI. It clarifies it. The best applications are the ones that make human judgement better informed and faster applied. A management team that can see emerging issues earlier, compare patterns more quickly and reduce time spent assembling information is in a stronger position than one relying on fragmented data and instinct alone. But the decision still belongs with the operator.
Good deployment is also selective. Not every workflow should be touched. Not every process benefits from automation. Some parts of a business are already working well and should be left alone. Others involve enough customer sensitivity or contextual complexity that a human-led process remains the right answer. The mistake is forcing the business into a model that looks efficient from the outside but weakens quality in practice.
AI is best treated as an operating tool, not an identity. Used well, it can materially improve speed, visibility and execution. Used badly, it becomes another layer of noise. The businesses that benefit most will be the ones that stay grounded: clear on where it helps, disciplined about where it does not, and focused on better decisions and stronger execution rather than fashion.
