Two ways forward, depending on your maturity: a hackathon to bring your teams on board and prototype in 1 to 3 days, or a turnkey implementation from framing to adoption. Always with cybersecurity built in from day one.
Start with the hackathon if you are looking for use cases. Go straight to implementation if the need is already clear.
AI Hackathon
For: leadership and teams who want to quickly find and test the right use cases.
Outcome: concrete prototypes, AI-literate teams and a prioritised roadmap.
See the formatsAI Implementation
For: an identified use case (invoice sorting, document assistant, customer service...) that needs to go to production.
Outcome: a system integrated with your tools, secured, measured and owned by your teams.
See the stepsDoing both? 20% of the hackathon price is credited if the implementation starts within 3 months.
AI Hackathon
An intensive, coached and secure format where your staff build AI solutions themselves on your real business challenges.
AI literacy and first no-code prototypes on 3 business challenges.
Deliverables
3 to 6 teams build working prototypes, presented to a leadership jury.
Deliverables
Multiple teams or sites, data prepared upfront, co-facilitation and internal communication.
Deliverables
D-15
Framing workshop (2 h)
Selection of 3 to 6 business challenges, data inventory, GDPR rules, sandbox preparation.
Day 1 - morning
Kick-off & AI literacy
What AI does well (and badly), risks, good practices, AI Act. Mixed teams are formed.
Day 1 - afternoon
Guided ideation & prototyping
Structured method, no-code / low-code tools, TrAInyng coaches inside the teams.
Day 2 - morning
Build & test
Prototypes tested on anonymised data, time saved measured, security check.
Day 2 - afternoon
Pitches & jury
5-minute demo per team, leadership + TrAInyng jury, selection of projects to industrialise.
D+7
Debrief
Written report, estimated gains, 90-day roadmap and implementation budget.
AI Implementation
A 4-step method with a decision point at each phase: you only commit the next budget once the previous step has proven its value.
Analysis of the process, the data and the expected gain. AI Act classification and GDPR impact assessment if needed.
Deliverables
One use case, 5 to 15 pilot users, indicators measured before and after.
Deliverables
Integrations (API, SSO), access management, logging, testing, documentation and user training.
Deliverables
Monitoring of answer quality and costs, model updates, support and quarterly review.
Deliverables
Typical use cases we run as a hackathon or an implementation.
Liège
Hundreds of supplier invoices keyed in manually every month.
Automatic extraction and pre-booking, compatible with the Peppol e-invoicing mandatory in Belgian B2B since 2026.
Goal : Cut keying time drastically, a human validates.
Brussels
Hours lost searching case law and internal templates.
Research assistant on the firm's document base, with cited sources and EU-hosted data.
Goal : Find the right document in seconds.
Antwerp
Overflowing order mailbox, in Dutch, French and English.
Sorting, extraction and assisted replies to emails, orders created in the ERP.
Goal : Handle simple requests the same day.
Ghent
Customer service overwhelmed during peak periods.
Multilingual conversational agent connected to order tracking, with human escalation.
Goal : Answer recurring questions 24/7.
Namur
Repetitive questions from beneficiaries and files to summarise.
Smart FAQ and file summaries, under a strict data protection policy.
Goal : Free up time for human support.
Illustrative examples: goals are set and measured during the framing of your project.
EU hosting preferred, vetted data processing agreements, enterprise plans with no model training on your data.
Risk classification, record of processing, user information and the AI literacy obligation (art. 4) covered.
Cybersecurity is our DNA: access management, logging, testing and a security review before go-live.
You own the code, the automations and the documentation. No vendor lock-in.
Our 70-20-10 approach: 70% practice, 20% exchange, 10% theory, so your staff truly adopt the tool.
If you do not yet know which processes to hand over to AI, start with the hackathon: your teams identify and test the use cases. If the need is already precise (for example automating invoice processing), go straight to implementation framing.
Yes. We work in a sandbox with test accounts, on anonymised or synthetic data. No real personal data is used without a legal basis and a data processing agreement.
No. Teams are mixed (business, support, management) and prototypes are built with no-code or low-code tools. Our coaches handle the technical side.
You receive a written report and a costed roadmap. If you start an implementation within 3 months, part of the hackathon price is credited.
Yes, as soon as your teams use AI at work. Article 4 has required a sufficient level of AI literacy since 2 February 2025, and authorities can enforce it since 2 August 2026. No certificate is required, but a documented training effort is: the hackathon includes this module, a certificate and the attendance list.
Yes, depending on your region and the provider's accreditation. In Wallonia, Start IA funds 50% of a framing mission (capped at €5,000) and Tremplin IA a proof of concept up to €20,000 (2026 calls closed, next ones to watch). In Flanders, the KMO-portefeuille covers 30% of training for small businesses. We check eligibility with you before quoting.
Prototypes by the end of the hackathon. For an implementation, the pilot delivers measured results within 3 to 6 weeks, before any larger investment decision.
Indicative prices excluding VAT (21%), venue and travel outside Belgium. Firm quote after the framing workshop.
15 minutes to frame your need and choose the right format. No commitment.
