< cd ..

2026-10-29· Kaizen Research

Introducing KAIZEN

Copilot, Cursor and Claude Code are already part of many developers’ daily routine. Some report big productivity gains, but in most companies the adoption is ad hoc and uncoordinated. One developer swears by an agent, the next one has never tried it, and nobody can say whether the team as a whole is getting faster or just moving effort from writing code to reviewing it.

KAIZEN is a research project that tries to close that gap.

The problem

Moving from personal experiments to a shared way of working raises questions that a single developer never has to answer:

  • Which tasks do we delegate to an agent, and which do we keep?
  • How do we protect architecture, quality and security when more code is written by agents?
  • Are the gains sustainable, or do they only shift the effort to reviewing and validating?
  • What are the rules for data protection and intellectual property?

Agents also need enough context, consistent quality rules and fast feedback loops to give reliable results. Without those, it is hard to scale up, and just as hard to measure whether it works.

What we are building

KAIZEN helps Flemish software companies adopt AI coding agents in a structured and responsible way. The expected results are:

  • an adoption framework and maturity model to assess readiness, find relevant use cases and grow from experimentation to broader adoption
  • playbooks for AI-assisted development, covering the development environment and the collaboration between developers and agents
  • demonstrators for representative use cases, such as specification-driven development and working with legacy codebases
  • training material: exercises, case studies, videos, articles, workshops and a guide for software entrepreneurs

How we work

The project combines applied research with direct input from software companies, in three workstreams:

  1. Understanding adoption needs through interviews and an analysis of current practices.
  2. Developing practical guidance on quality, security, tool integration, feedback loops and human-agent collaboration.
  3. Testing in realistic environments with demonstrators chosen together with an advisory group of companies.

Who it is for

Companies building software products, teams that want to go beyond individual experiments, and technology providers working on AI coding tools.

KAIZEN is carried out by Sirris with the University of Antwerp and Karel de Grote University of Applied Sciences, and is financed by VLAIO.

Get involved

The first user group meeting is on 26 November in Antwerp. See the event page for the agenda, and follow this site for updates as the framework, playbooks and demonstrators take shape.

$ cat ./partners

  • Sirris
  • Universiteit Antwerpen
  • Karel de Grote Hogeschool

$ cat ./funding

VLAIO

$ exit 0 — KAIZEN, a COOCK+ project funded by VLAIO