Scout

Scout starts with discovery. You find businesses in a geographic area, usually through a map, and save the ones worth investigating. From there it helps gather context on each one: what's on their website, whether they look like a fit for a particular service, and for UK companies, their registration details and signals about their financial health.

That research is organised around playbooks. A playbook is a predefined pitch for a specific service, setting out who the ideal client is and what checks to run before anyone reaches out. This is what keeps the process consistent. Rather than everyone qualifying leads their own way, the team works to the same definition of a good fit, and the whole pipeline stays visible instead of sitting in one person's inbox. When a lead looks promising, Scout helps draft the first contact for email, a call or LinkedIn based on what's been learned, before the opportunity moves into our main CRM.

Under the surface, Scout runs on Next.js. The part that does the real digging is an AI research assistant built on browser-use, which lets the AI work the live web the way a person would, investigating a specific business and saving what it finds as facts the whole team can see. That turns one-off research into something everyone benefits from later.

A lot of that work takes time, so we run it through Inngest, which handles background jobs. In practice that means the slow tasks, like the AI researching a company or pulling financial data, happen behind the scenes without anyone sitting and waiting for a page to finish. Building Scout for ourselves was also a way to work with the same AI and background-processing tools we use on client projects, on a problem we understand better than anyone. It's a fair reflection of how we build: pick the right tools for the job, and make something the people using it actually want to open each day.