Better systems. Better work.
People didn’t build their businesses to spend their days moving information between screens.
Yet a working day can disappear into chasing an update, entering the same information twice, or gathering the details needed to prepare a response. Those tasks are necessary. The effort they demand is worth questioning.
For us, the promise of AI is more room for the work that needs a person: deciding what happens next, noticing what a customer hasn’t said, or knowing when the usual answer isn’t the right one. A useful system should support that knowledge, not ask people to set it aside.
Start with the work, not the technology.
Before choosing a model, we want to understand a working day. Where does someone wait? What gets entered twice? Which exception always needs the same person? The team usually knows where the friction lives. Sometimes the answer is AI. Sometimes it’s conventional software or a simpler handoff. The right choice is the one that helps the work—not the one that makes the best demo.
Remove repetition. Preserve judgment.
Preparing a response isn’t the same as deciding to send it. A system can gather the facts and suggest a next step while a person remains responsible for the decision. We make that boundary explicit: what can happen automatically, what needs review, and how someone corrects or escalates a mistake. People should be able to question the system, not simply accept its answer.
The system should carry the repetition. People should keep the judgment.
For Hamel Gallery, we connected the catalogue, collector inquiries, and fulfillment records in one operating platform. The point was to support the work around the collection—not to substitute software for the gallery’s knowledge of the art.
Explore the Hamel Gallery projectBuild for the day after the demo.
The missing attachment. The request nobody expected. The service that goes offline. These aren’t distractions from the design; they’re part of it. We build with checks, testing, and recovery paths suited to the operation. Reliability takes work, and no system is beyond failure. When something breaks, the team needs to see what happened and know what to do next.
Design for the people who use it.
A tool can work perfectly and still make someone’s day harder. Another login, another queue, another unexplained result. We involve the people who will use it, listen to what doesn’t fit, and make room for training and adjustment. People need to understand what the system does and when it needs them. Adoption is part of the work, not something to hope for afterward.
Leave the team with ownership.
The system belongs in your business, not in a black box only its builder understands. Documentation, training, and a clear handoff help the team know what it can change and where it needs support. There will still be maintenance, third-party services, and decisions to make. Ownership means understanding those responsibilities—and having a clear path forward when the business changes.
More room for judgment, craft, and relationships.
Tell us which part of the day keeps getting in the way of the work that matters. We’ll start there.

