Teams Need Better Judgment, Not Just More AI Tools
A team I know was frustrated that their reports kept missing things, so they bought a tool that could produce the reports faster. Six weeks in, the reports came out quicker and still missed the same things, now with cleaner formatting. The tool had done exactly what it promised. It just turned out the original problem was never speed. It was that nobody had agreed on what a good report needed to notice in the first place.
This happens more than anyone likes to admit. A tool is a clean, purchasable answer, and the discomfort it is meant to solve is usually something messier underneath.
Here is the uncomfortable part. Adding a capable tool to a judgment gap does not close the gap. It speeds up whatever judgment you already had, in whichever direction it was already pointing. A team with sharp judgment and a good tool gets faster and better. A team with shaky judgment and the same tool gets faster at being shaky, and now with the quiet authority of something that looks automated.
The tension worth naming
Part of why teams reach for tools is that tools are legible. You can put one in a budget, announce it, point to it. Judgment is none of those things. It is diffuse, it builds slowly, and you cannot screenshot it for a status update. So when something feels off, the tool is simply the thing you can actually buy, and buying it feels like progress even when the real gap sits somewhere a purchase order can’t reach.
A capable tool in the hands of unclear judgment mostly produces confident mistakes faster.
What reaching for a tool often signals
It is worth treating the urge to add a tool as information. Sometimes it points at a genuine bottleneck, and the tool is the right call. Often it points at something else: the team has not agreed on what good looks like, or who decides, or which trade-offs matter. A tool will happily automate around that disagreement without ever resolving it, which can feel like relief for a while and then quietly costs more.
None of this is an argument against tools. It is an argument for being honest about which problem you are actually solving before you solve a different one faster.
A more useful way to look at it
A simple question cuts through most of this. What decision is this tool meant to improve, and could the team make that decision well without it? If the honest answer is no, the team could not make the call well by hand, then a tool will not rescue the judgment. It will scale the confusion. The better sequence is to get clear on the decision first, even roughly, and then bring in a tool to make the good version faster.
This is slower than buying something, and less satisfying in the moment. It is also the difference between a team that uses AI to amplify good thinking and one that uses it to mass-produce mediocre thinking with a nicer font.
Worth sitting with
The last time we reached for an AI tool, were we solving a speed problem or a judgment problem we hadn’t named?
Have we actually agreed, as a team, on what a good version of this work needs to notice?
If we took the tool away tomorrow, could we make this decision well by hand?
Where might a tool be quietly automating around a disagreement we haven’t resolved?
Building shared judgment is slower than buying a tool, but it is the thing that makes the tools pay off, and it is worth being deliberate about. If you want to read more on holding the line between speed and good thinking, Strategies to Navigate AI’s Dual Promise of Opportunity and Risk is a useful next read.
If your team keeps reaching for tools and landing in the same place, it can help to work through what good actually looks like with someone outside the room. A focused session with a practitioner in Compass is often more useful here than another piece of software.
