AI analyses incoming work and proposes, a human decides
The problem isn’t the technology
AI can take over a large part of the work today, in processes as well as in software. Yet most organisations don’t get faster or better. The reason: AI is added on top of the old way of working, instead of redesigning the work itself.
The wrong use
AI is rolled out as a personal productivity tool, while processes, roles and architecture are still designed for a world without AI.
Too many hand-offs
Work is split across specialised roles and teams. Every hand-off loses context, and AI can’t recover what was never passed on.
No ownership
Speed without architecture and accountability produces processes and systems that get harder to change, not easier.
What we do
Two services you can use separately or combine, depending on your question.
AI First Process Design
Before choosing technology, we redesign the process itself, starting from your business goals and customer journeys, rather than bolting AI onto an existing process. Per task we decide what AI automates, where AI supports and where a human decides. You get a target process, priorities and a feasible roadmap.
AI First Software Delivery
Experienced professionals take responsibility for analysis, architecture and delivery, from first concept to daily use. AI agents support development and testing within fixed agreements, with automated quality gates and human validation. Everything is built on reusable building blocks.
Small team, full accountability
Where a traditional project passes work from role to role, we run one line from need to go-live.
Traditional project versus our approach
One point of contact, from idea to use
Experienced professionals combine business analysis, architecture and delivery. Whoever understands your process also builds it.
AI under control
AI agents help us build and test software. Each agent works within fixed agreements and has one task: designing, building, testing or reviewing. Automated quality gates check every change and cannot be bypassed. Our team of experienced professionals makes the decisions, safeguards quality and decides when the software is ready for use.
A foundation that already exists
We don’t start from a blank page but from reusable building blocks: data, permissions, workflow, screens. Your project starts with what makes your organisation different.
AI under clear governance and security. With access control, logging and human oversight. We also protect against prompt injection and against data leaks through an AI assistant.
Software for your value chain. From customer relationships and orders to production, logistics and billing. We connect these processes to your ERP and other systems. For example, AI can extract order details from emails and prepare a draft order.
Three ways to start
Vision and roadmap
2 to 6 weeks, depending on scope
Together we define where your organisation wants to go and how processes, software and AI contribute. You get a target picture, quick wins and a concrete project proposal.
The proposal can also be traditional software without AI.
Proof of concept
10 working days, over 2 to 3 weeks
We build something that works on your own data and show what it means for your business. Goals and success criteria are fixed in advance.
Fixed price, deducted from the first project.
AI Literacy
1 day, tailored to your organisation
A practical training in which your people learn what AI and agents can and cannot do, with attention to data protection, reliability and human control. It also helps you build the AI literacy that the EU AI Act (Article 4) asks of every organisation that uses AI.
A low-threshold way to get to know how we work.
Already have a clear scope? Then we can start a project straight away.
Three ways to start that all lead to a project: vision and roadmap, proof of concept or AI Literacy. With a clear scope, a project can also start straight away. Project and support and further development flow into each other in both directions.
Development version with demo data.
One person. From A to Z. Live in use.
For Raak Millegem, one person, Koen Renders, built a complete portal: membership management, registrations, a webshop and online payments, reporting and workflow, plus an AI assistant, an MCP integration and document recognition. With separate test, acceptance and production environments on European cloud.
Estimate without AI: comparable to a team of six people for two years, built up per activity (analysis, development, testing, deployment and documentation).
Experienced people who take responsibility themselves
Kasim El Bastani
Clients and vision
Co-founder of C-Clear Partners, which grew to around 85 people before being acquired by an international IT group. With 20+ years in consultancy, Kasim turns business goals into a clear vision and a roadmap that holds up. For two global technology groups he delivered a global vision for their future CRM platform in six weeks; for one of them including a detailed roadmap, AI opportunities and redesigned, AI-driven processes. He stays close to delivery and steps into projects himself when needed.
Master of Engineering in electronics and computer science (VUB) with 20+ years in highly regulated environments: government, public safety and banking. He designs architectures that are secure, feasible and maintainable, from integrations across vendors in segregated environments to the replacement of legacy systems. He started as a developer and knows the technology down to the code, from Java and REST to React and PostgreSQL.
Koen has over 25 years of experience in business applications, architecture and programme management. He developed software products at CIPAL and, as CEO of Green Valley Belgium, led the shift to product development. As Practice Lead at Ravago, he led 45 IT professionals who built software for finance departments worldwide. He guided acquired companies from process redesign to ERP go-live. He designed the robust platform on which the Raak Millegem portal runs. He translates business operations into software that connects departments and integrates with existing systems.