We build AI that keeps running in the operation.

Where we fit, and where something else fits better

Anyone in these sectors looking to get started with AI chooses, in practice, between three routes. We are one of them, and not always the right one. Below, for each route, is where it is good and where it stops, so you can see whether your question belongs with us or is answered faster and more cheaply elsewhere.

• The sector's awareness course

  Trade associations and public initiatives offer short, often free AI courses. evofenedex has run a one-day Practice Training AI in Logistics since June 2026, and De Nationale AI-Cursus (the National AI Course) is free to take online.

  Where it is good. Fast, low-threshold and doable for a whole team at once. Your people learn to work with general tools like ChatGPT and Copilot and get a shared vocabulary. If AI barely lives in your organisation yet, this is a solid first step, and we recommend it ourselves.

  Where it stops. It stays at tool proficiency. Nothing is built on your own data in your own work process, there is no support afterwards, and no classification or accountability within your standards comes out of it.

• The compliance document tool

  Online providers that generate an AI register, a policy document and a set of templates for a subscription or a one-off fee.

  Where it is good. Cheap and fast. If your question is mainly paperwork, for example because a client is asking for an AI policy, such a tool delivers that paperwork. For an organisation that barely uses AI in critical processes, that can genuinely be enough.

  Where it stops. A template does not know your process. It does not touch your shop floor, trains no one, and does not translate your AI use into the CCPs and standard points your auditor tests against. If you choose this route, budget separate time and money for that translation and for training.

• The generalist AI implementation agency

  Agencies that build AI automation and bring the team along, for whatever sector walks in.

  Where it is good. There are good builders here. They work broadly, know the technology, and many train the client seriously alongside the build. For an office environment or a sector without a heavy standards framework, that is often exactly enough.

  Where it stops. The translation to your sector stays with you. An agency that does not encounter HACCP, FSSC 22000, GMP+ or GlobalG.A.P. daily cannot classify your line's image recognition in the language the NVWA or your certifying body will use.

What we do is the combination these three routes leave apart: training your own people while we make one real work process work with them, and recording that work immediately in the language of your own standard. Not because compliance is the reason to start, but because an AI application in an operation governed by HACCP, FSSC 22000, BRC, IFS, GMP+ or GlobalG.A.P. only keeps standing once your QA lead and your auditor can read it. The accountability record is the evidence, not the product.

If your question is only awareness, start with one of the courses above. If your question is only paperwork, a document tool gets you there faster. We are the right choice when a process in your operation needs to start running that your own people carry on and that you can account for.

Where our knowledge runs deep

We are deliberately narrow. Three sectors, two connected fields of expertise, and the standards those sectors work within.

Operations in logistics, food manufacturing and agri-food

We work from the primary work process: inventory and demand forecasting, order processing and admin, route planning and delivery, quality control and waste, harvest and labour planning. Our case material and templates come from real Dutch operations. We work within the frameworks your sector already knows and your auditor tests: HACCP, FSSC 22000, BRC, IFS, GMP+, GlobalG.A.P., GDP and GDPR. AI does not change those frameworks; it has to fit inside them, and that is exactly how we set it up.

EU AI Act and AI literacy

Since 2 February 2025, Article 4 of the EU AI Regulation has asked organisations that use AI to take measures that support the AI literacy of their staff, appropriate to their role. The Digital Omnibus (Regulation (EU) 2026/1744, in force since 27 July 2026) softened that wording and explicitly states that no specific level of literacy per person needs to be guaranteed. From 2 August 2026 the Regulation applies generally, with key supervisory roles for the Dutch Data Protection Authority and the RDI in Dutch supervision. We translate that regulation into what you can actually show: trained people, an adopted AI usage policy, and an accountability record recording who was trained, in what and when. We firmly distance ourselves from providers who threaten fines that do not hold up legally. We do not give legal advice; what we deliver are drafts and documentation for your own lawyer to review.

Our model: training and implementation together, with the right senior

Two choices shape how the work looks with us. The first: training and implementation sit in the same track, with the same team. A stand-alone course leaves knowledge behind without a working process, an advisory project leaves a working process behind without the knowledge to carry it on, and either way it stalls within six months. We do both at once, on your own work process.

The second: we do not work with a fixed roster of generalists who also do every sector, but with a curated network of senior practitioners, and we assemble the team that fits your sector and your question for each assignment. A food producer gets someone who knows the plant floor and HACCP from the inside. A 3PL gets someone who understands planning, driving times and margins. A greenhouse grower gets someone who understands climate data and yield. A fixed team is by definition a compromise, because it cannot hold the deepest knowledge in every sector. This way you do not pay for overhead that does not match your question, and you do not get a junior who has to learn your sector on the job.

The bar we select on

Anyone who stands in front of your operation on our behalf meets a fixed selection bar. We look at proven practice, not a polished slide deck.

• Demonstrable implementation experience in logistics, food manufacturing or agri-food, earned in real operations rather than only in the classroom.
• Seniority in operations or supply chain, so the conversation with your planners, QA leads and board is at the right level.
• Working knowledge of the sector's quality and compliance frameworks, from HACCP and FSSC 22000 to GDP and GDPR.
• The ability to transfer knowledge, so your team carries it on themselves. We train people who run it in their own hands.

What binds every assignment, regardless of who is on it, is one method, one quality standard and one principle: AI advises, the human decides and signs. Our templates, workflows and case material are built around the real work processes of these sectors, so the specialist who comes to you does not start from zero but from a proven base.

How a Practice Day and a track actually run

Our approach rests on one idea: in operational sectors, AI only delivers value when it sits inside the real work process, with human end-responsibility preserved. Starting small beats planning big.

• Practice Scan

  In a few questions we map which work process in your operation is the best first candidate for AI. Free, and evidence-based.

• Practice Day

  One concrete work process, at your own site, with your own team and data. By the end a working workflow is running, your team is trained, and the accountability record is ready.

• Ongoing guidance

  The heart of the partnership. Monthly sparring, reviewing workflows, adding new use cases and staying in control.

• Scaling, bespoke

  From single processes to AI across the whole operation, always tailored.

AI output is a draft. A human checks it, decides, and remains responsible. We design human approval on the critical steps explicitly into the process, rather than trusting that someone will pay attention. What the system proposes and what a person approves is logged, with role-based access, so you can show afterwards who approved what and when. That audit trail is the difference between AI you can account for and AI that stays a black box.

What you can hold us to

We hold ourselves to a fixed editorial and substantive standard, in our publications, our documents and at the table. This is that promise, made explicit.

• Substantiated, or not stated. We do not invent cases or figures. We cite market figures only with a source. Where we describe an illustrative approach, we call it illustrative.
• Sober, no hype. No buzzwords and no promises of fixed savings percentages up front. What is achievable depends on your assortment, your data quality and your process, and we say so.
• No scaremongering about fines. We are honest about what the law does and does not require. Panic stories about direct fines for the absence of a training are not grounded in the law, and that is how we describe them.
• Demonstrability over certificates. The AI Regulation prescribes no mandatory certificate and no fixed number of hours. What counts is that you can show what you have arranged. That is what we steer on.
• No legal certainty claimed. We do not give legal advice. Legal values and documents are drafts for you to have reviewed by a lawyer, and we mark them as such.

This is the same standard our publications on the insights page follow. You can read our content as a sample of how we work: substantiated, sourced, without sales talk.

In short

• Within your standardsHACCP, FSSC 22000, BRC, IFS, GMP+, GlobalG.A.P., GDP and GDPR.
• Hosted in the EUWebsite, portal and database run in a data centre in the EU, with an audit log and role-based access. The external services we use are listed in the privacy policy.
• The human decides and signsManual approval on critical steps, human end-responsibility built in.
• Substantiated contentNo invented cases or figures, market figures only with a source.
• Aligned with the EU AI ActArticle 4 demonstrably arranged, connected to your existing quality system.

What we base our work on

We build our approach on public, authoritative sources. A selection you can check yourself:

• Regulation (EU) 2024/1689 (EU AI Act), EUR-Lex
• Regulation (EU) 2026/1744 (Digital Omnibus on AI), in force since 27 July 2026, EUR-Lex
• Dutch Data Protection Authority, EU AI Act
• Dutch government, supervision of the European AI rules
• CBS (Statistics Netherlands), use of AI technology by Dutch companies
• CBS, companies use AI most often for marketing or sales (12 December 2025): 17 percent nationally, 16 percent in manufacturing and 7 percent in transport and storage, preliminary 2025 figures
• evofenedex, trade association for trade and logistics: published research on AI adoption in the sector
• evofenedex launches one-day Practice Training AI in Logistics (3 June 2026)
• De Nationale AI-Cursus, free online AI basics course, a collaboration of multiple organisations coordinated by Stichting Lowercase