Designing BenchSwap around the experiment

When someone books a place to stay for a week, they already know what they need. A research team planning a short sprint of experiments can arrive at two very different stages: one researcher already has a finished protocol and knows the exact biosafety level and add-on instruments to ask for, while another has an early hypothesis—or a computational model that needs wet-lab validation—and wants a few hours with a specialist to scope the experiment before choosing gear. Even with a defined protocol, a photo of an open bench doesn’t tell you whether a host’s instruments have the exact measurement capability an assay requires.

To make flexible access work for both groups, I designed our website intake around the project rather than static property listings. Researchers with a defined protocol can specify their space and equipment add-ons directly, while founders with an earlier idea can start with rough dates and an incomplete list—with a separate registration path as we build out our specialist network.

Rather than pretending a multi-part lab setup can clear instant self-checkout, the live website captures what the team wants to accomplish and hands it to us to confirm scientific fit, equipment availability, safety requirements, and host terms. That gave real prospects an easy starting point immediately while we learned from every inquiry.

Modular lab access also means separating when someone is at the bench from what their experiment occupies the rest of the week. A team running a three-week study might only need a workstation two days a week, while its cells occupy an incubator and its samples take up minus-80 freezer space every single day—alongside add-on instruments like a shaker or plate reader needed for specific steps.

I built continuing storage and modular equipment choices directly into our intake questions and configurator prototypes. Those prototypes tested how a team could combine shared instruments, dedicated add-on gear, and equipment they brought themselves—even when the bench space and individual instruments ran on different schedules.

Walking through those prototypes by hand with my operations cofounder also exposed places where rigid software rules undercut that flexibility. During one test, I caught a defect that prevented a user from selecting individual workstations without first renting an entire multi-seat bench. I directed the coding agents to fix the selection and pricing logic so a small team could book only the seats and add-ons its project actually needed.

Using AI coding agents let me turn product ideas into working software in days without waiting to hire an engineering team. I wrote the specifications, set the visual direction, tested every workflow from the customer, host, and operator seats, and directed fixes for misleading price displays, mislabeled safety questions, and signup friction.

Our earliest full-stack prototypes mapped out the complete on-demand vision—browsing spaces, configuring modular equipment, exploring specialist support, and stepping through agreements using fictional listings. After spinning out a preview for a real prospective host lab and seeing in customer conversations how much hands-on coordination went into matching an experiment to equipment and host terms, I streamlined the public website into the fast, guided project-request flow live today.

With a two-person team, helping our first customer get into a working lab mattered more than maintaining a complex automated platform. Prototyping ahead with AI let us test the modular vision, launch a practical front door for real inquiries, and see clearly which recurring coordination steps are worth automating next.