State of play, 22 September 2026. The Plan IA360 was presented on 21 September. There are no governing rules for the business voucher, no call for applications, and no application deadline or form to apply through. The plan places a controlled pilot in the first half of 2027 and the general call for applications before the end of that year, dates that depend on its regulatory development. Nothing that follows is a procedure: it's the calendar of what your business can prepare in the meantime, and of what's better left held back for now.
The Plan IA360 announces that the roll-out of the AI business voucher will incorporate a prior maturity diagnosis, identification of a specific use case, and subsequent measurement of the impact on productivity, and that its design would include conditions on training staff and embedding AI into internal processes. The analysis part of that method (define, measure, specify) can start being applied today. What you should not do just as freely is contract or carry out, now, the part of the project you would want the voucher to fund, because nobody can yet confirm whether it will count.
Why the order matters: the incentive effect
In many public grant schemes, only what starts after applying for the aid is eligible for funding. If a business has already paid for, contracted or carried out an action before a call for applications exists to apply to, that action — and in some schemes the whole project — risks being left out even if it fits perfectly with the criteria published afterwards. This is known as the incentive effect: the aid has to tip a decision that hasn't been taken yet, not reward one that was already taken independently.
The Plan IA360 has no governing rules yet, so nobody can confirm whether the business voucher will follow that same rule. Nor can it be ruled out. A timeline that encourages having the proof of concept and the pilot already carried out before a call for applications exists could leave the business unable to get funding for the part of the project it has already paid for out of its own pocket.
What to prepare, and what to leave held back for now
The plan sets only two dates of its own for the voucher: the controlled pilot in the «primer semestre de 2027» ("first half of 2027") and the general call for applications «antes de final de 2027» ("before the end of 2027"). Around those two points there's a boundary worth respecting: defining, specifying and measuring puts nothing at risk under the incentive effect; contracting and executing does.
| When | What to prepare (define, specify, measure) | What not to execute yet |
|---|---|---|
| Now, no deadline | Define one or two candidate use cases and measure their baseline: time, errors and cost of the process as it works today | Contract the development or integration of the chosen use case |
| Between now and the end of 2026 | Specify the proof of concept: what will be tested, with what data and with what measurable success criteria | Execute the proof of concept with a provider on the assumption that "it will already count" (the plan also excludes the mere subscription to licences) |
| First half of 2027 (the voucher's controlled pilot) | Design the training plan for the people who will use the system, with its own content and schedule | Contract or deliver that training now in the expectation that the voucher will recognise it afterwards |
| Before the end of 2027 (general call for applications) | Have the diagnosis, the use case and the baseline documented, ready in case the governing rules ask you to evidence them | Execute the part of the project you want to submit to the voucher before governing rules exist |
Before contracting or executing the part you'd want funded, wait for the governing rules: in many grant schemes only what starts after the application is eligible, and the IA360 voucher's rules don't exist yet.
What can be advanced with less exposure
None of the above means stopping. The maturity diagnosis, defining the use case and measuring the baseline are analysis work, not roll-out work, and that's why they expose you less to the risk of the incentive effect than contracting or executing does. But the plan includes the prior maturity diagnosis within the roll-out of the voucher itself, and without governing rules nobody can guarantee whether one done before the call for applications will count the same way: if you commission it now from an outside provider, plan for the business to pay for it. What is never lost, in any case, is its usefulness. It's the work covered by Summum IA's AI strategy consulting service: a maturity diagnosis, identification and prioritisation of use cases by ROI and risk, and an 18-month roadmap, lasting two to four weeks, in sessions and with a deliverable. The result is useful with or without the voucher, and only once governing rules exist will it be time to decide whether it's worth waiting for the call for applications or carrying on without it.
If the project is already under way
Some businesses will already have started, before knowing about the plan, a project that would fit the voucher today. For those cases there's no general answer: it depends on what the call for applications ultimately says about spending prior to the application, and that won't be known until governing rules exist. The prudent approach meanwhile is not to speed up or extend that execution on the assumption that the voucher will cover it, and to document dates, costs and scope from now on, in case the call ends up admitting actions already under way.
Frequently asked questions
Can I already start executing my AI project, or do I have to wait for governing rules?
It depends on which part. The diagnosis, defining the use case and measuring the baseline are the part that exposes you least and can be done already, although the plan ties the diagnosis to the voucher's own roll-out: if it's commissioned outside, plan for the business to pay for it. Contracting or executing the part you'd want funded is indeed better left until governing rules exist, because of the incentive effect.
What is the incentive effect, and why does it affect me if there's no call for applications yet?
It's the rule, common in public grant schemes, that only what starts after applying for the aid gets funded. Since the IA360 voucher's governing rules don't exist, nobody can confirm whether it will apply here or on what terms, and that's why the risk of executing ahead of time is real, not an exaggerated precaution.
When will it be known whether my project fits the voucher?
Not before governing rules are published. The plan places a controlled pilot in the first half of 2027, before the general call for applications, which would arrive before the end of that year. Until then, neither the amount per business nor the eligibility criteria are public.
Is it worth documenting the work if in the end there's no voucher, or it arrives different from what's expected?
Yes. The diagnosis, the defined use case and the measured baseline are useful whether the project gets funded or not, because it's the work any well-planned AI project needs so as not to stay just an intention.
The rest of the method the plan describes is in the guide to the Plan IA360, and the detail of what type of project the voucher would fund is in what AI projects fit the IA360 voucher.
Flagship project 11 (AI business voucher), Diffusion area, Plan IA360, La Moncloa, 21 September 2026. Full official document (PDF).
Last verified: 22 September 2026.