Practical AI-agent workflows with testing and human oversight

Dynamic Devices brings hands-on experience of integrating AI agents into software and embedded-product engineering workflows.

Through our Briar and engineering-harness work, we are developing approaches that combine persistent project knowledge, coordinated tasks, automated checks and explicit human approval. Our focus is helping people retain context and control as AI becomes part of everyday work.

We can contribute practical experience in structuring agent-assisted development, maintaining project knowledge across sessions, connecting agents to development tools and defining testing and review checkpoints.

We distinguish working implementations from capabilities still under development. External validation remains an important next step, particularly around reliability, oversight effort and measurable business benefits.

We would welcome collaboration with industry partners bringing a defined workflow, and with specialists in AI evaluation, security and human factors. For a BridgeAI consortium, we could contribute implementation experience and help design a focused evaluation comparing AI-assisted workflows with existing practice.