Case studies

Agent-led, Human-gated SDLC with Simitive

Simitive builds workforce planning and workload modelling software for universities, healthcare and professional services, with AI at every stage of how it ships. We advised them on the architecture for moving coding agents off laptops and into the delivery pipeline.

Simitive
Client
Simitive, a UK software company
Engagement
Software lifecycle advisory
Advised on
Cloud coding agents, AI workflows, CI/CD, test environments

Who Simitive are

A UK software company whose workforce planning and workload modelling software helps universities, hospitals and professional services firms plan the people and hours behind their work.

How we helped their SDLC

Coding agents had been running in supervised sessions on developers' laptops, one workstream each. We advised the team on the architecture for cloud agents covering requirements analysis, code review, and robust evaluations, so they could build better agentic systems themselves.

The software lifecycle, before and with agents
  1. Ticket

    Before

    Analysed when someone remembers to run it

    With agents

    Rated and given test cases the moment it is created

  2. Refine

    Before

    Starts from the raw backlog

    With agents

    Starts from a rated, pre-screened queue

  3. Build

    Before

    One supervised local session per developer

    With agents

    Cloud agents pick up tickets and build unattended, several at once

  4. Review

    Before

    Runs when a developer tags it

    With agents

    Runs on every branch by pipeline rule

  • a person
  • an agent
  • an agent proposes, a person decides

What we can do for you

If you need consulting help with your software development life cycle, including where agents should fit in and where human input is necessary for robust results, contact us.

“Alex’s support was instrumental in ratifying our processes and tools as we moved to an agentic development lifecycle.”
Natasha Bennett · Simitive