Case Study: How a Harrow Manufacturer Replaced an Unsupported Scheduling System and Cut Ongoing Costs by Around 80%

Executive Summary

The client is a manufacturer based in Harrow, Greater London, running production and job scheduling on a bespoke system built years earlier by an external developer. That developer had since stopped actively supporting the software — no updates, no roadmap, no meaningful support — while the business was still paying a substantial licence fee every year simply to retain access to it.

Sysgraft built a replacement: a bespoke, AI-powered production scheduling platform connected via API, designed specifically around how the client actually plans and runs jobs on the shop floor.

We are deliberately not publishing the underlying pound figures here — the comparison holds regardless of the specific numbers, and every engagement is scoped and priced individually. What matters is the shape of it: a legacy system with an open-ended, ever-present annual cost and no ongoing investment from its developer, replaced by a fixed-scope build with a transparent, much smaller monthly fee.

The Problem

Production scheduling is not a peripheral system for a manufacturer — it is how jobs get planned, sequenced, and tracked on the shop floor. The client's existing system had been built bespoke for them some years prior. It worked, in the sense that it still ran. But the developer who built it had moved on to other things and was no longer meaningfully supporting it: no feature development, no real troubleshooting, effectively no relationship beyond an invoice.

That left the business in an uncomfortable position familiar to a lot of manufacturers running older bespoke or niche software:

• Paying a significant annual sum for continued access to a system that was not improving and was not going to improve.
• No leverage. With one developer holding all the knowledge of how the system worked, there was no real negotiating position and no competitive pressure to keep the price honest.
• Ageing technology underneath a system that staff had learned to work around rather than one built to make their jobs easier.
• No path to improvement. Any request for a new feature or a fix went into a queue with no committed timeline — if it was answered at all.

The trigger for change was straightforward: paying that much, every year, indefinitely, for a system with no future, made less sense every year that passed.

The Solution

Sysgraft ran a discovery sprint on-site to understand exactly how scheduling actually worked day to day — not just what the old system technically did, but how the team used it, worked around it, and what genuinely mattered for planning jobs accurately.

The result was a bespoke, API-connected production scheduling platform with AI built into the core of how jobs get planned — helping surface scheduling conflicts, suggest sequencing, and reduce the manual effort that used to go into building and rebuilding the schedule by hand. Being purpose-built for this client, rather than a generic off-the-shelf scheduling tool, meant it could be shaped around their specific production process from day one.

Critically, this was a fixed-price build against an agreed scope — not an open-ended relationship with an hourly or annual dependency. Once built, ongoing hosting, maintenance, and support moved to a transparent monthly subscription, agreed upfront, with a fixed term.

The Outcome

The most immediate, measurable change was cost: moving from an indefinite annual licence fee to a one-off build plus a fixed monthly subscription cut ongoing software costs by roughly 80%. Even accounting for the one-off cost of discovery and build in year one, total first-year spend still came in well under half of what a single year on the old system had cost.

Beyond the cost comparison, the business now owns a system built specifically around how it actually schedules production, with an active support relationship behind it rather than a single developer who had effectively stopped engaging. Future changes and improvements are now something that can actually be requested and delivered, rather than a queue that never moved.