How to fund an ERP, a CRM or an AI agent with EU funding

An ERP that keeps stock and orders up to date, a CRM where no customer slips through the cracks, or an AI agent that answers the phone when your team is busy: under certain conditions, all of these can form part of an EU-funded digitalisation project. The question isn’t just “is it possible?”, but “how do I describe the project so it is eligible and, above all, useful after the funding ends?”.

In the guide below you will see how an ERP, a CRM or an AI agent fits into EU funding, which costs you can include, which costs stay with you and how to build the budget. The example rules come from the SME digitalisation guides published in Romania in 2026, but every call has its own rules, so always check the guide for the current call.

What the 2026 digitalisation guides say about ERP, CRM and AI

The clearest recent example is call 2, “Digital enterprises for an advanced economy” (“Întreprinderi digitale pentru o economie avansată”), under the Centre Region Programme (Programul Regiunea Centru). The official call sheet lists among the eligible activities:

• adopting digital technologies to improve internal processes: management of production and raw materials, service quality, customer interaction, logistics, data management and analysis systems;
• developing simple AI solutions, such as chatbots or recommendation engines, and complex AI solutions;
• digital communication with suppliers and customers;
• digital maturity audits and staff training.

The version of the guide put out for public consultation in April 2026 explicitly named ERP, CRM, PLM and RPA systems and accepted software licences, including temporary subscription-based licences, for a fixed period. The announced grant is between €15,000 and €200,000 per project, as de minimis aid, and intangible assets are mandatory in the project.

Other regional programmes use similar, but not identical, wording. So the first step is not choosing a supplier, but reading the list of eligible costs in your region’s guide, together with a consultant.

How each type of system fits in

ERP: running the business internally

An ERP links orders, stock, production, invoicing and reporting. In the language of the guides, it usually falls under “digital technologies to improve internal processes”. A custom ERP can be classed as software development or as a licence, depending on how the supplier offers it and what the guide accepts. What matters is that it is clear what the business receives: modules, users, integrations, documentation, training.

CRM: customer relationships

A CRM keeps track of customers, quotes and communication. It usually falls under “customer interaction” or “digital communication with suppliers and customers”. If the CRM automatically receives enquiries from your website, WhatsApp or phone, it is worth describing those integrations too, because they show the real change in process.

AI agent: chatbot or voice agent

A chatbot or an AI voice agent that answers customers and logs requests straight into the system can fall under “AI solutions” where the guide explicitly provides for them, as in the Centre Region. At Neurotech, agents run on Claude, the model developed by Anthropic, and data is stored in Supabase databases on servers in the EU. Technical details like these help the assessor understand what you are buying and how the data is protected.

Sample budget: what is included and what stays with you

Let’s take a patisserie with several outlets that wants to stop losing phone orders and keep its stock and orders in one place. A realistic project might include:

The example is not random: a chain of patisseries in Sibiu we work with uses an AI voice agent and a custom ERP and recovers around 20 calls a day that used to be lost orders. That project was not EU-funded, but it shows the kind of result worth describing in an application: a measurable problem and a concrete change.

What almost always stays with you

• Co-financing, set by each guide.
• Ineligible costs and VAT, according to the rules of the call.
• Cash until reimbursement: you often pay first and recover the grant later.
• Costs after the project: hosting, usage and maintenance. These depend on volume (how many calls, how many messages, how many users) and have to fit your monthly budget.

How to build the project, step by step

1. Start from a measurable problem. “We lose calls at peak times”, “we do stocktakes by hand”, “quotes get lost in email”. Guides generally reward projects that change processes, not just purchases.
2. Audit your processes. Where is data entered twice? Where does the customer wait? Some calls even fund the digital maturity audit itself.
3. Ask for a technical quote broken down by module. A good quote clearly separates development, licences, integrations, training and maintenance. That way your consultant can put each line in the right category.
4. Check the procurement rules. The guide tells you how to choose the supplier and which documents you need. Don’t assume; ask your consultant.
5. Plan the implementation in stages. Each stage should have a verifiable deliverable: a working module, a tested integration, completed training. This matters both for reporting to the funder and for you.
6. Think about the day after. Who will run the system? What will it cost each month? What happens to the data if you change supplier?

What to ask your supplier before you apply

Your consultant needs clear technical information to write a coherent application. Before you submit, ask your technology supplier for:

• a quote broken down by module, with separate prices for development, licences, integrations, training and maintenance;
• a description of the solution an assessor can understand: what problem it solves, who uses it, what data it processes;
• an implementation timeline with verifiable deliverables for each stage;
• information on where the data is stored and how it is protected, especially if the system processes customer data;
• an estimate of monthly costs after the project, based on your real volume.

You can find details on how we handle data on our data security page.

Common mistakes

• A shopping list instead of a project. An ERP, a CRM and a chatbot that don’t talk to each other change nothing in the business.
• Software chosen for points, not for use. If your staff don’t use it, you are left with the reporting obligations and none of the benefits.
• A vague quote. “Complete AI system – €50,000” can be assessed neither by the consultant nor by the authority.
• Spending too early. Check in the guide from what date costs become eligible before you sign any contract with a supplier.
• Ignoring the de minimis ceiling. De minimis aid for a single undertaking is capped at €300,000 over three years, under Regulation (EU) 2023/2831. Aid you have already received counts towards it.

How we work on projects like these

Neurotech AI is a technology supplier. We don’t write funding applications and we don’t guarantee that a project will be approved. We work alongside your EU funding consultant: we prepare the technical quote by module, the description of the solution and the implementation plan, then we build the system and put it into operation.

Our public implementation prices start from €2,500 for a custom ERP or CRM, €1,500 for an AI voice agent, €800 for a WhatsApp chatbot and €1,200 for a process automation, excluding VAT, paid in instalments tied to deliverables. Monthly costs depend on volume and we discuss them for your specific case.

If you want to see what an ERP, a CRM or an AI agent would look like for your business before the call opens, book a free 30-minute call. You will come away with a clear idea of what you need and what to take to your consultant.