The EIC Accelerator funding (grant and equity, with blended financing option) by the European Commission (EC) and European Innovation Council (EIC) awards up to €2.5 million in grant and €10 million in equity financing per project (€12.5 million total) and is designed for startups and Small- and Medium-Sized Enterprises (SME), often supported by professional writers, freelancers or consultants.

This article is Part 2 of the series and explores the use of Artificial Intelligence (AI) in EIC Accelerator grant proposal evaluations (see ChatEIC). Part 1 can be found here.

Why The EIC Should Move Fast

The integration of AI into the evaluation system provides very distinct advantages to the EIC, but it needs to move fast to proceed with the experimentation phase before the number of submissions skyrockets. The EIC should start earlier rather than later since it will not have the luxury to integrate a half-baked product, as this could destroy its entire evaluation process (and reputation). It must gradually test and adjust a new AI process and only push it to the real evaluation system once it has reached a high quality threshold.

Luckily for the EIC, this is quite easy to execute since it has access to decades of proposals, evaluations, and funding decisions. This means it can simply lean on its past data to rapidly train and improve new AI systems. Once a new system has been conceived, it can be tested against real data, and evaluators can be used for reinforcement learning by giving their feedback on the outcomes.

Of course, since the luck factor has been present for a long time, the AI might end up confused as to why one project was rejected while another was funded. However, there is still an opportunity to speed up developments.

Speed and Scale

The EIC likes to talk about scalability, but its evaluation process is anything but. More submissions? Let's hire more evaluators. More interviews? Let's hire more EIC jury members. More technology challenges? Let's hire more program managers.

That is not a scalable process. Yes, one can do unscalable things in the inception phase, as Paul Graham advises, but that will become infeasible as user numbers grow. A scalable process has asymmetric benefits and not a linear relationship between the team size and the results. If application numbers grow, the EIC is forced to match them linearly by adding new evaluators. This linear relationship is its lynchpin.

It is also forced to continuously ensure that the quality remains high while simultaneously diluting the entire process - a fool's errand.

The more jury members and evaluators the EIC hires, the harder it will be to manage them since they will be more difficult to oversee and brief. Having a small team to communicate with about what you truly care about is much easier than a hall full of people, half of whom aren’t even listening. It will also be difficult to keep the quality of evaluators sufficiently high if the numbers have to grow fast.

Turning the remote evaluators into jury members would also be a disastrous strategy, as any past EIC Accelerator applicant who has read through many iterations of evaluator comments will attest to.

Switching the EIC Accelerator evaluation process from human-based to AI-driven is not simply a high-risk challenge; it is a high-reward opportunity—something the EIC should embrace if it had any understanding of the word irony.

Precise Instructions vs. Subjective Briefings

The core of the opportunity is quite simple. The EIC would be able to hardcode its instructions directly into the AI and would not be subject to a kilometers-wide Telephone Game (or Chinese Whispers) where evaluators apply not what the EIC wants but what the EIC probably wants.

There is a lossiness to the current approach, and this has already been known since 2021 when EIC members announced on LinkedIn that applicants would have a legitimate complaint only if evaluators made a mistake by misunderstanding the EIC’s rules in their evaluation. And such mistakes happened frequently.

Albeit, the EIC later backtracked on the rebuttal process and removed its support system entirely. You want to reach out to the EIC? Good luck. Read the FAQ.

However, the AI approach delivers a key benefit since the EIC leadership team can introduce top-down changes within a few minutes by simply adjusting the instructions. They can sit in on the Step 3 EIC Accelerator juries, read the Step 2 business plans, and perform real-time adjustments to ensure the quality of the applications remains high.

The EIC A(I)-Team

The EIC can set up an AI team whose sole purpose is to design prompts, manage the IT system, and integrate the feedback from the EIC Board. Since the EIC is a bureaucratic organization and not a tech company, this has the additional benefit of requiring no AI researchers — it will be enough to have experienced prompt jockeys on board.

Instead of briefing a room full of people on a Zoom call where bored evaluators are told to ensure it is really deep DeepTech and also aligned with lofty policy buzzwords, the EIC Board can have an intimate meeting about the AI instructions, adjust the overall criteria, and assess edge cases.

From then onwards, the EIC could actually make progress with every deadline since the evaluations will become more aligned with the EIC’s desires and will lead to measurable improvements.

Key Performance Failures: The Simple Things In Life

Here is the best indicator to prove that the EIC has failed to increase the quality of its evaluation process and has, in fact, objectively made it worse: the EIC Accelerator Step 3 interview success rates.

The Step 3 rates are not subjective and are not artifacts of the increasing applicant numbers. They are a failure of the EIC’s leadership. Step 1 and Step 2 of the EIC Accelerator process are designed to filter companies and ensure that only the best end up in the Step 3 interviews, but the rates have fallen.

The interview success rates used to be 50% back in 2021 and had fallen to only 20% by 2023 (see Smack My Pitch Up). In 2025, these rates have fallen to only 16% (see Success Rates Updated). Considering that the Step 3 EIC Accelerator interviewees do not receive travel reimbursements anymore means that for every 100 interviewees, 84 have to go home with a loss and a reduced cash balance.

The EIC has continually failed to improve its evaluation process and has let it sink from 50% to only 16%. To be fair, one could argue that many more excellent companies now reach the interview and that the budget is too limited to fund them all, but that is a simple excuse to avoid the fact that Step 1 and Step 2 are designed to filter applicants and avoid bottlenecks.

That they have failed to do.

The State Of The Step 3 Interviews

With evaluators turning into jury members, the jury roster increasing rapidly, and the number of interviewees rising, it is not surprising that the EIC is struggling to find the best companies. The evaluation process has more band-aids than lasting solutions at this point. Removing travel reimbursements for applicants is likewise a sign that things will get worse in the future and not better. Money flows more steadily to evaluators and consultants instead of applicants.

If a sudden surge of amazing DeepTech companies has emerged from the woodworks over the past years then it would be more plausible that the interviews have just become more competitive, but it is highly unlikely. Based on experience, there are still many less fitting companies that go to the interview, with many consultants already knowing that they will probably not get funded. This is the case now more than ever, unfortunately.

With AI-driven evaluations, the EIC finally has the opportunity to fix its most important metric: the EIC Accelerator Step 3 interview success rate. If the rate is high and the EIC Board is happy with the quality of applicants, then the evaluation process was a great success. If the success rate is low and the jury members complain about poor quality, then it was a failure.

With AI, the EIC can fine-tune in real-time.

And Here Is The Greatest Part

The EIC takes a comfortable 6 to 8 weeks (or more) to assess Step 1 and Step 2 proposals. This is due to the limitation of the human context window (i.e., the brain) and the difficulty in coordinating the evaluations of thousands of proposals with 3 to 4 evaluations each.

But AI does not work like that. The EIC can have the evaluation done in just a day. And then? They can have experts review some key proposals and then repeat the evaluation with enhanced instructions. Instead of waiting for 8 weeks until the evaluators have finally found a consensus, the EIC can spend 4 weeks iterating AI evaluations to get the roster of interviewees just right.

This process can be as powerful as conducting 50 evaluations instead of 1 to get the best companies to reach the interview. It is often difficult to wrap one's head around the power of AI tools, but the speed of improvement can be groundbreaking, and the EIC could fix its most important metrics in less than a year after letting them hit rock bottom.

Additionally, the EIC's AI team, EIC Board, and other management staff can coordinate for a month to fine-tune the selection themselves, powered by AI, and only release the results once they have a high degree of certainty that they actually found the best. And they would still be twice as fast as human evaluators with 10 times the quality enhancement.

Conclusion: Another Criteria Bites The Dust

AI represents a major opportunity for the EIC since it can turn its declining evaluation process around and turn it into a best-in-class proposal evaluation system. All it needs to do is to turn some of its staff into IT and AI prompting experts and to let a contractor build the (simple) architecture. As long as the EIC will pay that contractor, the system will be a blessing for the applicants and jury members alike.

But how will the EIC criteria look like if AI has to do the job? What about gender, diversity, policy, and buzzwords? That will be discussed in Part 3.

These tips are not only useful for European startups, professional writers, consultants and Small and Medium-Sized Enterprises (SME) but are generally recommended when writing a business plan or investor documents.

Deadlines: Post-Horizon 2020, the EIC Accelerator accepts Step 1 submissions now while the deadlines for the full applications (Step 2) under Horizon Europe are:

The Step 1 applications must be submitted weeks in advance of Step 2. The next EIC Accelerator cut-off for Step 2 (full proposal) can be found here. After Brexit, UK companies can still apply to the EIC Accelerator under Horizon Europe albeit with non-dilutive grant applications only - thereby excluding equity-financing.

Contact: You can reach out to us via this contact form to work with a professional consultant.

AI Grant Writer: ChatEIC is a fully automated EIC Accelerator grant proposal writer: Get it here.

EU, UK & US Startups: Alternative financing options for EU, UK and US innovation startups are the EIC Pathfinder (combining Future and Emerging Technologies - FET Open & FET Proactive) with €4M per project, Thematic Priorities, European Innovation Partnerships (EIP), Innovate UK with £3M (for UK-companies only) as well as the Small Business Innovation Research (SBIR) and Small Business Technology Transfer (STTR) grants with $1M (for US-companies only).

Any more questions? View the Frequently Asked Questions (FAQ) section.

Want to see all articles? They can be found here.

For Updates: Join this Newsletter!