RIA vs IA vs CSA, which Horizon Europe action fits you

Here is the part most guides skip: you do not choose the action type. The call does.

Every Horizon Europe Pillar 2 topic is published as a RIA, an IA or a CSA, and you apply under that type, full stop. So "RIA vs IA" is never a decision you make inside a proposal. It is two decisions you make around it: which topics suit your project's maturity and nature in the first place, and how the action type changes what you must emphasise once you write.

Get either wrong and the failure mode is predictable. A consortium chases an IA topic with a technology that is nowhere near demonstration. Or it wins the right topic and then writes a RIA as if it were an IA, promising markets instead of proving methodology. Both are common, and both are avoidable. This page shows you how to read the type, and how to write to it. For the full method behind the proposal itself, start with how to write a winning Horizon Europe proposal.

The three action types at a glance

The TRL ranges are typical and indicative, not eligibility rules. The funding rates and templates are firm. Now the detail, because the differences that lose points live in the detail.

RIA: research and innovation action

A RIA exists to generate new knowledge or a new technology, or to explore whether something is feasible at all. Maturity is lower, typically an indicative TRL 2 to 5. Funding is 100% of eligible costs for every participant, which is why RIAs remain the natural home for universities and research organisations.

The Part B is the standard three-section template: Excellence, Impact, Implementation, with a 40-page limit in the current 2026 form (45 pages for lump-sum funding). The full structure is broken down in the Part B template.

What evaluators reward most in a RIA: scientific excellence, novelty genuinely beyond the state of the art, and a sound methodology. Impact is scored, but it is not weighted. That asymmetry should shape your effort: a RIA lives or dies first on the strength of its science.

IA: innovation action

An IA exists to produce plans and designs for new or improved products, processes or services, and to take them through first application: prototyping, testing, demonstration, piloting, market replication. Maturity is higher and closer to market, typically an indicative TRL 5 to 8.

The funding rate is the structural difference companies feel first: 70% of eligible costs for profit-making entities, while non-profit legal entities such as universities and research organisations still receive 100%. Budget your consortium with that split in mind from day one.

The Part B uses the same three-section template as a RIA, with the same page limit (40 pages, 45 for lump sum). The difference is not the form, it is the emphasis, and one scoring rule: for an IA, the Impact criterion is weighted 1.5 for the ranking (not for the thresholds). Evaluators reward a credible route to market and real exploitation, alongside the science. How that weighting works in practice is covered in how proposals are scored, and the impact pathway it amplifies is the subject of the Impact section.

CSA: coordination and support action

A CSA is not research. It funds accompanying measures: coordination and networking, studies, roadmaps, standardisation, awareness and dissemination, policy support, community building. No TRL applies, because there is no technology to mature. Funding is 100% of eligible costs.

The Part B is a different, shorter template. Its section 1 is "Objectives" plus "coordination and/or support measures and methodology". There is no "ambition" subsection, no state of the art, no TRL, because a CSA does no direct research. The page limit is shorter than a RIA or IA, in the region of 25 to 30 pages, set by the call's own application form.

What evaluators reward most: the value and reach of the coordination or support to the target community, and a credible plan to deliver it. A CSA is judged on service to a community, not on discovery.

How to tell which type a call is

You never have to guess. The topic states its action type (RIA, IA or CSA) in the work programme, in black and white.

But do not stop at the label. Read the expected outcomes, because they tell you what kind of project the Commission is actually buying. Research-heavy outcomes point to a RIA. Deployment or market outcomes point to an IA. Coordination or policy outcomes point to a CSA. If your project's centre of gravity does not match the outcomes, the topic is a poor fit even if the acronym looks right, and evaluators will feel the mismatch on every page.

Same facts, different story: how to write to each

The action type is not just an admin field. It is a reading instruction for your evaluators, and your proposal should follow it.

• In a RIA, lead with scientific novelty and methodology. Show precisely where you go beyond the state of the art and how your method holds up. Market claims are supporting cast, not the headline.
• In an IA, lead with the exploitation route, the market and the demonstration. Remember that Impact carries extra weight in the ranking: it is your strongest scoring lever, so build the Impact section like it decides the outcome, because for close rankings it can.
• In a CSA, lead with the value of the support to the community and the reach of your coordination. Do not pad it with a research-style state of the art it does not need: the template removed those sections for a reason.

And the three traps, each one a self-inflicted wound:

1. Over-claiming market impact in a RIA. It reads as not understanding the instrument.
2. Under-claiming exploitation in an IA. With Impact weighted 1.5 in the ranking, you leave your best scoring lever unused.
3. Writing a CSA like a research project. TRL tables and state-of-the-art reviews it never asked for signal that you did not read the template.

Match the type, then write to it

The action type question resolves into a simple sequence. Check the label in the work programme. Check that the expected outcomes match your project's maturity and nature, with the TRL ranges as an indicative guide, not a gate. Then write to the type: science first in a RIA, exploitation first in an IA, community value first in a CSA, on the right template with the right page limit.

The three-section structure itself, and exactly how the CSA version differs, is covered in the Part B template. The wider method, from call choice to submission, is in how to write a winning Horizon Europe proposal.

→ GrantForge's pre-submission evaluation scores your proposal against the criteria for its action type, so you see how it reads before the evaluators do.

Part of our guide on how to write a winning Horizon Europe proposal.