The most interesting reactions so far are not about AI. They’re about decision-making. People aren’t telling me their models are broken. They’re telling me that “We don’t agree on what the outputs mean.” That’s the real problem.
What makes this particularly hard to fix is that interpretive drift feels like disagreement about strategy, when it's actually disagreement about meaning, and those require completely different interventions. Most organisations respond by adding more process or clearer governance, but those tools assume people are working from the same semantic foundation, which is precisely what's missing. The interpretation layer isn't just ungoverned; it's often invisible to the people inside it.
Thank you 🙏🏻 Kyla. This is exactly it, especially the point about organisations treating this as a strategy problem rather than a meaning problem. Most interventions (process, governance, controls) assume shared interpretation already exists. Which is why they often make things look more aligned… without actually resolving the underlying divergence.
The most interesting reactions so far are not about AI. They’re about decision-making. People aren’t telling me their models are broken. They’re telling me that “We don’t agree on what the outputs mean.” That’s the real problem.
I wondered if you are seeing this too?
What makes this particularly hard to fix is that interpretive drift feels like disagreement about strategy, when it's actually disagreement about meaning, and those require completely different interventions. Most organisations respond by adding more process or clearer governance, but those tools assume people are working from the same semantic foundation, which is precisely what's missing. The interpretation layer isn't just ungoverned; it's often invisible to the people inside it.
Thank you 🙏🏻 Kyla. This is exactly it, especially the point about organisations treating this as a strategy problem rather than a meaning problem. Most interventions (process, governance, controls) assume shared interpretation already exists. Which is why they often make things look more aligned… without actually resolving the underlying divergence.