# CQFDev: From audit to AI > We audit your processes before writing code. Custom AI solutions, automation, and strategic consulting, based in Brussels, Belgium. AI Consulting & Development. We audit your business processes before developing your AI solution. [Schedule a call](https://calendly.com/francoisdevigne-dev/consultation) ## What we do **Analyst-first. End-to-end.** Concrete deliverables at each stage, built around your actual situation. ### AI Audit Structured review of your **processes, data quality, and team maturity** before any code is written. We identify where AI creates real value and where it adds unnecessary complexity. You leave with a **prioritised action plan** tailored to your constraints, delivered by the same team that will build the solution. ### Custom AI Solutions Full-cycle development of **LLM agents, ML models, and automated workflows** tailored to your context. Each system connects to your existing infrastructure via API. No vendor lock-in: **you own the code, the data, and the operational knowledge** from day one. ### Process Automation We rebuild your most time-consuming manual processes as intelligent, monitored workflows using **n8n, Python, and custom API integrations**. Documented, monitored, built with explicit failure handling. Your team **recovers capacity** for decisions that require genuine expertise. ## Our products **Products born from the field.** Each solution was born from a real need on the ground. ### Newsletter Automation *Industry body · Deployed* AI pipeline monitoring dozens of sources daily, filtering by relevance, and auto-drafting newsletter content for a European industry federation. Source monitoring, entity extraction, relevance scoring, and draft generation run fully automated. The editorial team reviews and approves. Editorial prep time reduced by over 70%. ### DonAction *NGO sector · In development* ML-powered donor analytics segmenting donors by giving behaviour, forecasting churn risk at the individual level, and surfacing targeted engagement recommendations. No data science team required, GDPR-compliant from the ground up. ## About CQFDev **Analyst mindset. Developer precision.** CQFDev is an AI consultancy and development studio founded in 2024 in Brussels, Belgium. We combine business analysis with hands-on engineering, and have already delivered production systems for a European industry federation and the NGO sector. No consultants who hand off to developers. No context lost between strategy and execution. One team takes you from the initial diagnosis to a running system in production. **François Devigne**, Founder & AI Developer · Brussels Business engineer (MBA) and former functional analyst on secure software projects across financial services and the public sector. Before founding CQFDev, François co-founded a Brussels cybersecurity firm, building its analytical infrastructure from the ground up. That background shapes CQFDev's analyst-first methodology directly. Brussels-based, GDPR-compliant by design. [View LinkedIn profile](https://www.linkedin.com/in/francoisdevigne/) ### The approach 1. **Audit**: Process inventory, data quality assessment, and team readiness evaluation. 2. **Identify**: We pinpoint where AI can have real impact, and where it cannot. 3. **Build**: Custom solution, built and deployed into your stack. 4. **Iterate**: Continuous refinement based on real-world performance. ### The expertise LLM Agents · ML Pipelines · Process Automation · Python · GCP · n8n · GDPR · EU Institutions · Data Analytics · API Integration ## Common questions ### What does "analyst-first" mean in practice? The first conversation is never about technology. We spend time understanding your processes before we open a text editor. In practice, this means a structured process inventory: what happens step by step, who touches the data, where decisions are made manually, and what constraints exist in the current system. This approach routinely surfaces constraints that would have derailed a solution built on assumptions: a legacy database format, a team without labelled training data, a compliance requirement that rules out a cloud component entirely. ### What types of organisations do you work with? Brussels-based SMEs, NGOs, European institutions, and industry federations. Organisations with real operational complexity that want AI solutions designed around their actual environment, not generic templates. In practice, this means organisations that have already tried to deploy off-the-shelf tools and hit the ceiling, or that need to justify an AI investment to a board or procurement committee and need a structured diagnostic before committing to a build. If you are at an early stage, the audit is the right first step. ### Do you build from scratch or integrate with existing tools? Both. We use proven frameworks like n8n, Python, and GCP, and connect them to what you already have. The decision on whether to build, configure, or integrate is made during the audit phase, based on your actual data, your team's capacity to maintain the system, and the complexity of the target use case. We avoid unnecessary reinvention and never introduce proprietary lock-in: the systems we build are designed to be understandable and maintainable without us after delivery. ### How is our data handled? All work is covered by NDA and handled GDPR-compliant. By default, your data stays in your environment. Cloud processing is only introduced when explicitly required, and every data flow is documented upfront before any development begins. For EU institutions and regulated sectors, we map every data movement, apply data minimisation at the architecture level, and document the legal basis for each processing activity. We can work entirely within your infrastructure if required: our tooling supports on-premise and private cloud deployments. ### How do we get started? A 30-minute call. We listen, ask the right questions, and give you an honest assessment of whether AI is the right solution for your problem. No pitch deck, no sales funnel. If the answer is yes, we outline what an initial AI Audit would cover and what it would cost. If the answer is no, or not yet, we will tell you that directly, along with what would need to change before AI becomes viable. We have no interest in building something that will not deliver. ## Get in touch **Let's talk about your project.** A problem to solve, an idea to explore? Describe it in a few words, we'll get back to you quickly. - Schedule a call: https://calendly.com/francoisdevigne-dev/consultation - Contact form: https://cqfdev.com/#contact - LinkedIn: https://www.linkedin.com/company/cqf-dev We read every message and reply personally. ## Languages - English: https://cqfdev.com/ (Markdown version: https://cqfdev.com/index.html.md) - Français: https://cqfdev.com/fr/ (Markdown version: https://cqfdev.com/fr/index.html.md) - Nederlands: https://cqfdev.com/nl/ (Markdown version: https://cqfdev.com/nl/index.html.md) --- © 2026 CQFDev. Brussels, Belgium. [Privacy Policy](https://cqfdev.com/privacy/) · [llms.txt](https://cqfdev.com/llms.txt)