AI Adoption
The Drawer Full of Tools
By John J. BakerJuly 23, 20266 min read
I spent an hour last week with a general contractor who is sharper on technology than most people I meet who sell it. He had tried everything. Estimating add-ons, an AI notetaker, a proposal generator, a scheduling assistant, a couple of chatbots. He was not a skeptic and he was not behind. If anything he was ahead, and he was frustrated in a way that only someone who has actually tried can be.
His line stuck with me. "Every one of these gets me about seventy-five percent of the way, and then it just stops." He was not complaining that the tools were bad. He was complaining that they were almost good, over and over, and that after a year of buying them nothing in his business had really changed except that he now had more logins and more tabs open.
That is not a technology problem. That is tool fatigue, and it is quietly becoming the most common condition I see in businesses that took AI seriously.
What tool fatigue actually looks like
The story is always some version of his. You read about a tool, or it shows up bundled in software you already pay for. It looks good in the demo. You buy it, you wire up one thing, it does that one thing about three-quarters of the way, and you tell yourself you will come back and finish the rest. You never do, because the next tool has already shown up.
Four tools, and nothing actually moved
Each one is genuinely good. Each one stops at about the same place.
Stops before your read that the number is light because the site access is bad.
Stops before deciding which of the twenty notes is the one thing that has to happen next.
Stops before the client who needs a phone call, not one more clean PDF.
Stops before the trade-off between two jobs that only the PM can actually make.
And they do not talk to each other. Four logins, four half-answers, nothing that adds up.
The problem was never the tools. It was buying them one at a time.
No plan coming in, so the tools arrive disconnected, each solving a slice nobody mapped, each stopping before the part that mattered. Add enough of them and you do not get a system. You get a drawer.
Look at what that picture is really saying. Each tool is genuinely good. Not one of them is the villain. The problem is that they arrived one at a time, each aimed at a slice nobody had mapped, and each one stopped at the exact point where the work started to matter. That last stretch, the part every tool leaves for you, is the reason you still feel like you are doing the job by hand.
And they do not connect. Four tools that each know a quarter of your process, none of them talking to each other, is not a system. It is a drawer. The fear of missing out fills that drawer fast, and a full drawer feels like progress right up until you notice the work is exactly where it was.
Why the last stretch is always the hard part
There is a reason it is always the last quarter that stops the tool, and it is worth understanding, because it tells you where the real work is.
A tool is built for the general version of your problem. It knows what an estimate is, what a proposal is, what a follow-up is, in the abstract, across ten thousand companies. What it does not know is your site, your client, your crew, and the judgment your estimator applies without noticing he is applying it. The first three-quarters is the general part, and a tool does the general part well. The last quarter is the part that is specific to you, and it was never written down anywhere the tool could read it. That is the same gap I have written about as the last slice a loop cannot reach, and no amount of tool-shopping closes it, because the missing piece is not for sale.
So adding a fifth tool does not help. It gives you a fifth thing that is seventy-five percent of the way to a slice you still have not mapped. The instinct when the current tool disappoints is to go find a better one. The better move is to stop shopping and get clear on the work.
The four things to be clear on before any tool
Here is the counterintuitive part, and the part the contractor and I actually spent our hour on. The way out of the drawer is not another tool. It is getting clear, on paper, before you touch anything, about four things.
The problem
What are you actually doing?
Name the one outcome this is meant to move. Not "use AI," not "be more efficient." The quote that goes cold, the report you rebuild every Monday, the follow-up that never happens.
The people
Who is involved in it?
Every hand that touches the work and the one that owns the outcome. The estimator, the PM, the office manager, the client. A tool dropped between two people who never spoke will not fix the handoff.
The process
What are the steps?
The real sequence, in order, including the messy handoff nobody wrote down. You cannot automate a process you cannot yet describe, and most of the value is in describing it honestly.
The judgment
Where does a human decide?
The calls that carry your name and the ones that need a person in the loop on purpose. Mark them before you build, so the tool accelerates the work instead of quietly taking over the part you meant to keep.
The tool is not one of the four. It comes last, and once these are clear, it comes easily.
Problem, people, process, judgment. Get those in focus and the technology has something solid to layer onto. Skip them and you are back at the drawer, adding a fifth tool to a pile of four.
None of that is technical. It is the work of describing your own business honestly, which is harder and more valuable than it sounds, because most of it lives in people's heads and has never been said out loud in order. What are you actually doing. Who is involved. What are the steps. Where does a human decide. Answer those four and you have done the thing almost everyone skips, and you have done the thing that makes every tool afterward land.
Notice the order matters, and notice what is missing. The tool is not one of the four. It is deliberately last, because a tool is only ever an accelerator, and an accelerator pointed at a process you cannot describe just gets you to the confusion faster. This is the same reason the best input still beats the best tool: the technology forgives sloppy thinking right up until it multiplies it.
Think about the whole business, not one problem at a time
The other trap in the drawer is that every tool solves exactly one problem, so you start thinking in single problems too. This tool for quotes. That tool for scheduling. Another for notes. But your business does not run as a set of isolated problems. It runs as a system, where a problem in estimating shows up as a margin problem in accounting, and a handoff between two people is where the whole thing actually breaks.
So the mapping is not just for the one task in front of you. It is for how the organization actually works. You are looking at the process, the people, and the technology together, and you are asking where the friction really is, which is almost never inside a single step and almost always in the seams between them. Map one problem well and you will help one task. Map how the work moves through your business and you will see which few problems, fixed in the right order, change everything downstream. That is the difference between buying tools and deciding what to actually fix first.
The judgment step deserves its own emphasis here, because it is the one people rush. Deciding where a human stays in the loop is not a limitation you accept because the tool is not good enough yet. It is a choice you make on purpose. Some of that last quarter should stay in human hands permanently, no matter how capable the technology gets, and knowing which parts those are is as important as knowing what to automate. Mark those seams before you build, or the tool will quietly take over the part you meant to keep, and you will find out from a client.
The fastest way there is slowly
I know how this sounds to a busy owner. You have a business to run and I am telling you to slow down and draw maps. But this is the part I most want to land.
The fastest way there is slowly
Same amount of time. The path that starts slower is the one that keeps going.
Tool-first. Buys, jumps to three-quarters, plateaus. The next tool does not raise the line.
Clarity-first. Slow through the mapping, then it crosses over and keeps climbing as the intelligence improves.
The tool-first path feels fast because it is fast, at first. You buy something and you get a jump the same afternoon. Then it plateaus, and the next purchase does not raise the line, and a year later you are the contractor with a drawer full of tools and a business that looks the same.
The clarity-first path feels slow, because for a while it is. You are mapping instead of buying, and it feels like falling behind while everyone else posts about their new tools. Then something happens that the tool-first path never gets to. The line crosses over and keeps climbing, because now the intelligence has something solid to accelerate, and the intelligence keeps getting better on its own. Every model that ships next just makes a well-mapped process faster. The same models pointed at a tangle just make a faster tangle.
That is why I keep saying the fastest way to go somewhere is slowly. The clarity is the part that compounds. The tools are the part that changes every quarter. Build on the thing that compounds, and you win no matter what launches next. This is the whole reason we tell owners to invest in the foundation before the tools: the tools were never going to be the durable part.
Where ClearOak comes in
If you are the contractor I described, ahead rather than behind, a drawer full of good tools and a nagging sense that nothing moved, you do not have a technology problem. You have an unmapped one, and that is genuinely good news, because it is fixable and it is fixable by you.
What we do is sit in that first, unglamorous stretch with you. We watch how the work actually moves, we name the four things on paper, we find the two or three seams where fixing the process changes the most downstream, and only then do we talk about tools, usually fewer of them than you already own. The point is never to add to the drawer. It is to make the drawer unnecessary.
If any of this sounds like your last year, that is the conversation to have. Schedule a call and tell me which tools you bought and where each one stopped. That map is halfway to the answer.
Prefer email? Reach me directly at john@clearoakconsulting.com and tell me the one that got you seventy-five percent of the way. The missing quarter is usually the whole conversation.