A Custom Jira Alternative for a 20-Person Software Team

The challenge

They weren't starting from nothing. They were already on Jira — and that's the part most people get wrong about custom software. The problem usually isn't "we have no system." It's "we have a system that doesn't fit."

For a 20-person software team, Jira brought a large surface area they never touched, while still missing the three things they actually ran the business on:

• Capacity per person. Who is genuinely available next sprint, in real working days — not abstract points.
• Epic management shaped their way, rather than reshaping how they worked to suit the tool.
• Manday-based change requests. This is the commercial mechanism that decides what gets billed. Tracking it outside the task system meant maintaining two sources of truth and reconciling them by hand.

That last one is the tell. When the thing your revenue depends on lives in a spreadsheet beside your project tool, the tool isn't doing its job.

What we built

A task and project management system modelled on the parts of Jira that worked, with the parts they never used left out — and the gaps filled in properly.

• Per-person capacity tracking — availability in mandays, visible before commitments are made rather than after.
• Epic management structured around their delivery process.
• Manday CR management — change requests tracked as first-class records inside the same system as the work itself, so scope and billing reconcile automatically.
• An MCP server. The system exposes a Model Context Protocol interface, so any AI agent can read and update tasks, capacity, and change requests directly. Standing up a project, checking who has room next sprint, or drafting a CR can happen through an agent instead of a browser tab.

The MCP layer is the part worth pausing on. Most task tools treat AI as a bolted-on summarisation feature. Exposing the system over MCP means the AI agents a team already uses can operate it as a tool — which is a different proposition from a chatbot that describes your backlog back to you.

The results

The system went live 3 months from kickoff and replaced Jira for the full 20-person team.

The measure of success here isn't a percentage — it's that capacity planning and change-request billing now happen in the same place as the work, instead of being reconciled by hand across two systems. And because they own the code, the tool can keep changing as their delivery process does.

Would this suit you?

Custom is not automatically the right answer. If your process is genuinely standard, off-the-shelf usually wins, and we'll say so.

It becomes the right answer when a tool forces you to work around it in the one area that matters commercially — as it did here. If that sounds familiar, get a free consultation and we'll tell you honestly which way we'd go.