Governance is a word I’m hearing a lot more. Sometimes, in the technical (ISO) sense, but more when discussing power dynamics.
On several recent deals, we’ve seen bright, young, and capable people in risk roles (within the proposed investee). Initially, that’s heartening. I’ve been mentoring folks earlier in their risk career for a few years, and one of the challenges is finding roles. However, as we get into the details around governance structures, independence, reporting lines, etc., you can see that these more junior people have all the risk responsibilities but no authority. Outsourced blame, in other words.
In this context, the person responsible for risk can:
🧐 Spot issues. 🧐 Raise concerns.
But they can’t:
🚩 Stop the people above them, because the structure doesn’t let them.
That’s not governance. That’s window dressing.
Thankfully, most of the investors I work with take board seats (sometimes observers, sometimes their own people, and often appointees at the board and exec level). One of the first tasks on the action plan (conditions precedent or subsequent to investment) is to create:
✅ Rules that apply upwards. ✅ Escalation paths that aren’t career-ending. ✅ Upskilling the board.
That last point is a subject of constant discussion, and will be a core focus in 2026 and beyond. What we’re learning is that boards are overwhelmed with data, but (sometimes) lack the know-how to (effectively and quickly) synthesise the data, and ask the right questions. For example:
📍 Training completion data is meaningless without insight into comprehension, user feedback, and indicators of whether it worked (reduced incidents, increased question…see next point). 📍 Reporting, grievance, and speak-up data requires context - too little indicates fear, apathy, or a lack of awareness/communication, too much indicates confusion around what to raise, weak/silent middle management, culture collapse, etc. 📍 Audit and other data analysis requires similar contextualisation to speak-up data - how much loss, seapage, write-off, absenteeism, etc., is ‘normal’, and so on.
That’s for starters.
The good news is that many of these issues can be resolved via a tech and human blend. Well-crafted ‘schemas’ can better aggregate, catalogue, synthesise, and surface anomalies (we’re doing this, at scale with a tool we plan to make open source, if trademarked, in the next quarter). But to read that data, act on it, and understand the Abraham Wald paradox requires a bit of knowledge transfer. Not as much as you might imagine, but asking questions is an art more than a science.
If you’re wondering who Abraham Wald was, he is responsible for the image below. During WWII, American bombers were suffering catastrophic losses. The military mapped the holes on returning aircraft (the image). The plan was to strengthen these areas until Wald pointed out that these were the planes returning. Those who never made it home likely suffered injury to the areas without the holes.
So, yes, give career opportunities to smart young people. But, when looking at counterparties and investees, remember that a risk function without upward power is just decoration.
Once the function is structurally supported, help the board spot the (visible and invisible) holes, and arm them with the power of better questions.