They Had a Tool for Everything. That Was the Whole Problem.

Picture someone on your team with a customer on the phone. To answer one ordinary question, they open the CRM in one tab, a couple of data tools in two more, email in a fourth, and the phone app in a fifth. By the time they've found the answer, the customer has already repeated themselves twice.

That was the scene at a financial advisory firm we started working with earlier this year. Not a broken business. A good one, actually. Sharp people, tools that all worked. That was exactly the problem: every tool worked, and there were far too many of them.

The request we almost always get

When a company feels that kind of friction, the ask that comes to us is nearly always the same. Build us one system. One place where everything lives. A single, beautiful dashboard that finally pulls it all together.

We get it. It sounds like the cure.

But a funny thing happens when you try to build the one-dashboard-to-rule-them-all. You usually end up with one more place to click. The seventh tab. Now people have six tools plus a dashboard that tries to mirror all six, and it's stale the second any one of them changes.

So we slowed down and asked a different question. Not "how do we show everything in one place," but "what are people actually trying to do when they open all those tabs?"

The pain wasn't missing information. It was the hunting.

The answer, almost always, was small. They wanted to look up where a customer stood. Check what had been said last time. Record a decision. See whether a recommendation had actually been acted on. Day to day, maybe eight or nine things made up the overwhelming majority of the work. The deep, occasional, hairy tasks were real, but rare.

That gap is the whole story. The information wasn't missing. It all existed, somewhere, in one of those tools. The pain was the hunting. Every common task meant navigating, clicking, logging in, and remembering which tool held which fact.

It's a tax you pay a hundred times a day and never see on any invoice.

One task made the point for us. To log a certain kind of recommendation and the conditions attached to it, a person had to walk through four separate manual steps across the system. Four steps, every single time, for something that happened constantly. Nobody enjoyed it, so sometimes it didn't get done. Which meant the firm couldn't even see the gap between what its people recommended and what customers actually did. The busywork wasn't just annoying. It was quietly erasing information the business needed.

So we built a smaller interface, not a bigger one

Here's the move we made, and it's the opposite of what we were asked for. Instead of building a bigger interface, we built a smaller one. Two parts.

First, a simple cockpit that pulls just the handful of daily feeds people actually use into one clean view. Not everything. The eighty percent. If someone needs one of the deep tools for a rare task, it's one click away, the full thing, untouched. We didn't try to rebuild those. We just stopped making people live inside all of them all day.

Second, and this is the part that felt like a glimpse of where things are going, a chat box. Instead of clicking through screens to find where a customer stands, you type the question the way you'd ask a colleague. "What did we last talk about with this account, and is anything still open?" The answer comes back in plain language, pulled live from the data underneath. No navigating. No remembering which tool. You ask, you get the answer.

We started calling it "interfaceless," half as a joke. But the joke was true. The best interface for most of the daily work turned out to be almost no interface at all. A question and an answer.

The CRM didn't go away, by the way. It got demoted, in a good way. It became the engine room, the place the data lives, and stopped being the thing a human has to wrestle with directly. People got the question box. The software kept the filing. If your CRM feels more like a chore than a help, that reframe is worth sitting with. We've written before about how most business software tracks what you sold, not who you know, which is a related flavor of the same disease.

The honest part, because a chat box over your data is easy to oversell

And we're not going to oversell it.

A question-and-answer layer is only ever as good as the data sitting under it. If the numbers are stale, a confident answer is just a confident wrong answer. So a big chunk of this build wasn't the slick chat part at all. It was making sure the data feeding it was live and trustworthy, with a reliable, constantly refreshed copy behind the scenes so the chat is never answering from yesterday's picture. Unglamorous plumbing. It's most of the job.

We also drew a hard line around anything that counts as professional judgment. In this firm, certain recommendations carry real weight. We were careful that the AI never quietly becomes the one making those calls. Where it drafts or surfaces something that sounds like advice, a human expert reviews it against their own standard before it goes anywhere near a customer. We spent real time stress-testing the system's wording against what the firm's most experienced person would actually say.

The machine handles the looking-up and the busywork. The human keeps the judgment. That line isn't a nice-to-have. In a business where being wrong has consequences, it's the whole ballgame.

And we locked the thing down: invite-only, password-protected, no open doors. When you put a plain-language window over a company's entire customer history, who can open that window matters as much as what it shows.

One last honest note. There's a real dependency in a build like this: getting clean, direct access to the data living inside someone else's software. That is not always a sure thing. We had to plan for the version where the easy door stays shut and we take the longer way around. Pretending those risks don't exist is how projects quietly fall apart. Naming them out loud, early, is how they don't. This is still a work in progress, not a victory lap, so you won't find any before-and-after numbers here. Anyone handing you tidy metrics this early is making them up.

What you can actually use

Here's the part that travels, whether you run a five-person shop or a fifty-person one. When your team feels buried in software, resist the urge to buy or build one more thing on top. More tools, even a unifying one, usually means more surface area, not less. Try this instead.

• Count the clicks. Pick one ordinary question a customer or a coworker asks you all the time. Now count how many tools, tabs, and logins it takes one of your people to answer it. If the answer is four or five, you don't have an information problem. You have a hunting problem, and hunting is pure waste. It's a cousin of the quieter cost we covered in how work stalls in the gaps between your steps.
• Separate the eighty from the twenty. Figure out the handful of things your team does every single day. Those are the ones worth making effortless, ideally one question away. The rare, complicated tasks can stay in the specialized tools where they belong. You don't have to rebuild everything. You just have to stop making people wade through everything to reach the common stuff.
• Let people ask instead of hunt. The strongest everyday use of AI inside a business right now isn't flashy. It's quietly turning "go find this across five systems" into "just ask, and get the answer." If you've ever wished you could ask your own data a plain-English question, that's the same instinct, scaled up to a whole stack of tools instead of a single spreadsheet.

A warning on that last one, though. This only works when people actually reach for it. A clever tool nobody opens is just an expensive bookmark, which is a trap we've watched plenty of companies fall into. That's a whole story of its own: why nobody uses the AI you bought. Build it around the real daily questions, not the impressive demo, and people will use it because it's genuinely faster than the old way.

The point

We didn't set out to prove a point about software philosophy. We set out to make one team's day less exhausting. But the lesson stuck, so we're passing it along with no strings attached.

The goal was never a prettier dashboard. It was giving people back the minutes, and the focus, that all that clicking was quietly eating. Sometimes the best thing you can build for someone is less.