Every team has its workarounds: the notebook next to the software, the overgrown Excel, the post-it stuck to the screen. These workarounds are the real map of the need. We start there.
The sign that a tool is wrong isn't the complaint in a meeting. It's the silent workaround: the one nobody mentions because it's part of the furniture. Spend a morning in your workplace and you'll see them within the hour: the notebook between the till and the kitchen, the shared file three people edit at once, the mobile app opened alongside a system that, in theory, already does the same job. A spec written in a meeting room misses these objects. A field visit finds them in under an hour.
Three workarounds we see almost always
Double entry. The data is typed into one system, then rekeyed into a second: either because the two don't talk, or because one is for doing and the other for proving. We've seen the same data typed four times in a day by three different people. Excel that drifts. It starts as a tracker. A year later it's the real control panel. Nobody decided this; it just happened. Paper kept alongside. The one that speaks loudest. It exists because the official tool doesn't understand what it records: a note, an oral instruction, an exception. The real craft is on that paper.
Why we come watch before proposing anything
A brief is built from what we're told and what we see. The two never fully match, not from dishonesty, from habit. You've lost awareness of the gestures you make every day. An outside eye picks them up. We watch four things: repeated gestures, interruptions, papers not on the computer, and silences: moments where someone waits without knowing why. Those four cover 80% of the real need.
What a theoretical spec misses
A need described in advance has been filtered by the person describing it. Three biases always: optimism ("it'll be fine"), generalization ("we always do it this way" when it's one case in ten), and hierarchy (what you write isn't necessarily what pinches the field team). We watch to bypass those three biases, not to catch you out, but to build a tool your team will actually use.
What we've seen on the ground
In a client restaurant, the chef kept a notebook next to the till to correct nightly discrepancies. That detail (invisible in a sales demo) shaped the entire real-time numbers app we built next.
FAQ
How long does the observation take? Between a half-day and two days, depending on complexity. Priced as a package, not by the hour, so you don't watch the meter.
Does it disrupt the team? Less than a briefing meeting. We're present, we don't speak unless spoken to, we take notes on paper.
Have a workaround that's been running for months? We'll come see what put it there. Then we decide together whether to build, replace, or leave alone. Let's talk →