Fred and I met at Oxford 13 years ago. We've lived in five houses together and were ushers at each other's weddings. In 2022, we started building software for the waste industry after seeing how poorly existing ERPs served day-to-day operations. We spent mornings at customer depots from 6AM, watching transport planners manually drag and drop jobs into routes. Plans took over an hour to build and were often obsolete within another hour. Our top customers told us they literally cannot train planners fast enough to keep up. The cognitive load burns most planners out within 18 months. We realised the problem wasn't missing features. It was that humans were being asked to do work they physically cannot do reliably. In January 2025, we pivoted to Dayjob: autonomous AI workers that plan and execute operational decisions in real time.

We started in October 2022, building ERP and inventory software for the waste industry, generating $150K in revenue. Two hard lessons emerged: large operators wouldn't migrate ERPs regardless of product quality, and better UI didn't solve the real problem of manual decision-making at scale. In January 2025, we pivoted to Dayjob, integrating with existing ERPs instead of replacing them and deploying AI workers that perform operational work directly. We started selling the product in August 2025. ARR has grown to $359k+ in seven months, with top customers achieving 8%+ efficiency gains from day one.

Industrial logistics relies on humans continuously re-optimising thousands of constraints in real time: vehicle capacity, weight limits, site access windows, driver hours, and regulatory rules, all while responding to hundreds of live job changes. A single day's route plan takes 90 minutes to build and is typically obsolete within an hour. Our top customers say they cannot train planners fast enough. Most burn out within 18 months. The problem is getting worse: experienced drivers who historically moved into planning roles are being replaced by office hires with no operational intuition, increasing errors and churn. ERP systems were built to digitise paperwork, not to make autonomous decisions. Operators are dependent on a role they can no longer hire or retain.

If Dayjob succeeds, industrial logistics companies won't have ops teams managing scheduling and dispatch. Dayjob will be the brain. Every route plan, real-time change, exception, and customer interaction will be handled autonomously. We start with transport planning, add real-time re-optimisation, then customer service, until the entire operational back office runs on Dayjob agents. The ERP stack doesn't disappear. It becomes AI-led rather than human-dependent.