Your organisation is AI-ready when the data, process, decision rights and know-how around one specific decision are in order. So start with a decision that recurs often, runs on reliable IFS data and has a measurable result. Your data does not have to be perfect. But who owns it, and who may approve an AI recommendation, must be settled before you begin.
In earlier blogs, Erry de Boer, CEO of Eqeep, shared his view of AI in finance, operations and sales and marketing. Before you dive in, you will want to know whether your organisation is ready. In this concluding blog, Erry looks at the reality of AI implementation.
Where do organisations really stand with AI?
AI use is growing fast, but more in marketing and sales than in production and service. Among Dutch companies with 50 to 250 employees, use rose from 20% in 2023 to 45% in 2025. These are provisional figures from Statistics Netherlands (CBS). Of all companies using AI, 35% apply it to marketing or sales and 18% to production or service processes.
Many companies are still hesitant to invest. They follow the trends but want proof first of how AI works with their existing software and in practice. That caution is understandable, because no one wants to invest in technology that does not fit daily reality.
Eurostat’s figures on AI in enterprises with ten or more employees show what holds companies back. In 2025, EU businesses that had considered AI but did not use it most often cited a lack of relevant expertise (71%). Next came unclear legal consequences (53%) and concerns about data protection and privacy (49%). Only 21% thought AI was not useful for their business.
So the obstacle is mainly know-how and clarity on rules and privacy. Below, you will see how IFS permissions limit the data AI uses, and what the AI Act expects.
Which data must be in order before you start with AI?
The data behind the first decision you want to hand to AI. For work orders at risk, that means asset status, open work orders and inspection plans. IFS Intelligence links those records to its answer, so you see where the figures come from.
Look critically at that data. Is it complete and reliable, are your records accurate, and is it current and consistent across all systems? It does not have to be 100% perfect, but without a good foundation AI becomes a wasted investment. The data cleansing does not have to cover your entire ERP system at once.
Those answers are only as reliable as your master data and permissions. Give each type of master data an owner, for example for assets, items and customers. You decide what quality is good enough to start.
How do you know whether your processes are ready for AI?
When you know how the work really runs, exceptions included. Check whether you have many exceptions or mainly work with standardised processes. IFS Process Intelligence uses your transaction data to show how work flows through IFS, with every deviation from the optimal path.
Those exceptions decide how much work AI can take over. A digital worker in IFS Loops runs a defined workflow and escalates to a person when judgement or approval is needed. If a process is full of exceptions, the work keeps coming back to your team.
Where data goes wrong between departments, and who owns a process, is covered in our article on gaining control of inefficient processes.
Who may act on an AI recommendation?
You settle that per decision, before the first use case goes live. Decide what AI may prepare, which actions need approval and how you will measure the result. For work orders at risk, you might measure how many are still completed on time.
In IFS Intelligence, Cowork prepares actions but only carries them out once someone has approved them. For work orders at risk, the planner might approve the prepared change. Underneath sits the Industrial AI Harness, which secures and governs the AI. It uses your existing identities, inherits permissions from connected systems and follows your approval steps.
AI only uses data it is authorised for. Your current permissions and approval steps in IFS become the rules for AI. Tidy them up first.
The AI features already in IFS Cloud, such as IFS.ai Copilot, are also switched on per role. According to the IFS documentation for release 26R1, users get Copilot through permission sets, bundles of access rights. For work orders at risk, the relevant sets include those for service managers, dispatchers and work order execution. Granting them is your application manager’s job.
How front-runners handle decision rights, with IT, Operations and Risk at one table, is set out in our insights from IFS Connect Benelux 2026.
Which roles do you need for a first AI use case?
Four. Put a name to each role, so every step from signal to decision has an owner.
- Process owner: accountable for the decision and its result, and decides what goes into daily use.
- Data owner: looks after the master data the decision runs on.
- Key users: test the use case against the work they do every day.
- Application manager: manages permissions and approval steps in IFS and switches AI features on per role.
In a mid-sized company, one person sometimes holds two roles. If so, write down who takes which decision.
What does the EU AI Act already require of you?
That you work on the AI literacy of your staff. Article 4 of the EU AI Act has required this of organisations using AI systems since 2 February 2025.
Since an amendment in July 2026, the law no longer prescribes a specific level. The duty to take measures remains, and national market surveillance authorities have supervised it since 2 August 2026. No certificate is needed. You can keep an internal record of the training and guidance you have given.
This also applies to simple use. If staff write advertising copy with ChatGPT, the European Commission says they should be informed about the specific risks, such as made-up answers (hallucinations). So, alongside your first use case, list the AI tools your people already use. Then set out per role what they need to know.
Where do you start with AI?
With one decision your team makes every day or every week, such as work orders at risk or late orders.
Organisations that have made a successful start with AI tackle the data and the process behind that one decision first. Only then come the applications we discussed in the previous blogs.
We see the same pattern in the winners of the IFS Awards 2026 in the Benelux. They start small, with one concrete process and a measurable outcome. They treat data quality as a project in its own right, started before the first model runs. We describe their approach in our insights from IFS Connect Benelux 2026.
We start with IFS Intelligence the same way, with your key users and your own data. You see the results and decide what goes into daily use. See how Eqeep starts with IFS Intelligence.
Sometimes the honest outcome is that AI is not the first step yet. If no decision recurs often enough, or the data behind it is missing, fixing the process or master data pays off first.
Questions about an AI-ready organisation
When is your company in the process industries ready for IFS Intelligence?
When the data behind one recurring decision, such as releasing a batch, sits in IFS Cloud or a connected system. If you still track batches and analyses on paper or in spreadsheets, IFS Intelligence is not your first step yet. Start with lot and batch tracking in IFS Cloud.
Which permissions should you tidy up first before switching AI on?
Those around your first decision: which roles may view the data behind it, who may record a change and who approves it. When a change falls outside a user’s permissions, IFS Intelligence prepares it and asks the responsible planner for approval. If a permission no longer fits the role, withdraw it before you start. How IFS Intelligence is secured is set out on our IFS Intelligence page.
Which IFS partner helps with the first IFS Intelligence use case?
Eqeep, IFS Specialized Partner in the Benelux. We choose the first decision with your team, check data, roles and approvals, and build a first use case with your key users. We work in Dutch and English, and our consultants also come on site in Belgium.
Which decision should AI support first in your organisation?
Tell us where you stand with IFS. After one conversation you will have a clear next step. Bring the decision you want AI to support first. Discuss your IFS question.
Stay one step ahead with IFS. Get release news and practical examples from IFS Specialized Partner Eqeep in your inbox. Free, and you can unsubscribe at any time.