Frequently Asked Questions

The EIC has removed rebuttal functions, preventing applicants from responding to evaluator comments and addressing inconsistencies. This shields the evaluation process from criticism and reinforces potentially flawed assessments.

The 'intelligence test' involves using the correct terminology (e.g., 'de-risk,' 'crowd-in'), understanding bureaucratic rules (e.g., TRL levels, budget requirements), and navigating potentially inconsistent feedback from evaluators.

A better approach would focus on removing generic content and streamlining the process to focus on essential information. AI could be used to identify key data points and assess the overall potential of the company, rather than grading the proposal as an essay.

If both applicants and EIC evaluators use AI to process the same information, it creates a redundant cycle. Applicants use AI to bloat concise data into a narrative, and evaluators might use AI to distill it back down, wasting time and potentially degrading the original information.

Evaluators often focus on minor details and adherence to a specific format rather than assessing the overall potential of the company and technology. This can lead to strong proposals from weaker companies succeeding while innovative companies with less polished writing fail.

The current process is bureaucratic and time-consuming, often taking 8-15 months with a low success rate. It prioritizes adherence to rules and buzzwords over genuine innovation, making it a poor use of resources for companies needing urgent funding.

Evaluators should prioritize assessing whether a company is impressive and a good fit for the EIC portfolio. This requires a more subjective approach, acknowledging that every company has flaws and focusing on overall potential rather than minor imperfections.

Using AI isn't technically plagiarism since it generates content from scratch, not by copying. However, AI learns from vast datasets, so the content isn't entirely original. It's a gray area, but currently, it's not considered plagiarism.

There are indications that the EIC might be experimenting with AI in Step 1 evaluations. Changes like typo-free reports and duplicate sentences suggest some level of automation in the evaluation process.

The EIC hasn't established clear guidelines or verification mechanisms regarding AI usage, so ethical considerations are currently sidelined. Applicants are free to use AI tools without penalty, especially since the evaluation process can be subjective.

The EIC Accelerator offers up to €2.5 million in grant funding and up to €10 million in equity financing per project. This blended financing option can provide a total of up to €12.5 million to support innovative startups and SMEs.

The EU should prioritize supporting scalable and disruptive business models, including IT, marketplaces, and B2C consumer apps. These types of companies can grow into industry giants, unlike smaller DeepTech companies that are often acquired or relocated.

Europeans need to learn to adapt to the changes brought about by AI developed in the US and China. Focusing on scalable business models and avoiding excessive regulation are crucial steps in responding effectively to the AI revolution.

The EIC initially launched an EIC Accelerator challenge focused on foundation models and LLMs. However, it was later downgraded to a generative AI challenge, likely due to the limited availability of advanced LLM development in Europe.

The EU faces challenges in AI development due to regulations, complex legal frameworks, and a lack of investment. These factors can hinder capital-intensive innovations, making it difficult for European companies to compete with US and Chinese AI developers.

Yes, the EIC launched an 'AI platform' in 2021, but it was shut down abruptly after two years due to a contract dispute with the IT service provider. This left many applicants in a difficult position right before a deadline.

Yes, the use of AI in EIC Accelerator grant writing is already prevalent, especially among applicants who don't use consultancies. The lengthy application process and specific content requirements make AI a valuable tool for generating high-quality content.

Include any IP information relevant to your company, such as a list of IP assets (patents, trademarks, domain names), technical graphics of discussed patents, legal analyses, patent applications, scientific publications, and other relevant IP documents.

Identify the innovative features of your technology and create keyword groups for each. For example, if your innovation is 'Green hydrogen generation from algae,' keywords could include 'Green hydrogen,' 'Hydrogen algae,' and 'Hydrogen biomass.' Use multiple keywords per innovation for a comprehensive search.

For each patent, list relevant metadata like the inventor, assignee, filing status, and territories. Then, explain how the patent differs from your innovation and why it doesn't pose an IP or commercial risk to your company.

The methodology section should explain how you prepared the FTO analysis, including the databases used (e.g., Espacenet) and the keywords chosen based on your technology's descriptions. Acknowledge that no FTO analysis is perfect and that the quality depends on the comprehensiveness of your keywords and research time.

Yes, it's recommended to state that you prepared the FTO analysis in-house and plan to contract an IP firm for a more thorough analysis during the EIC Accelerator project. This shows awareness of the limitations of your in-house analysis and a commitment to addressing IP concerns.

An FTO analysis assesses the existing IP landscape to ensure your company can patent its technology and commercialize its product without infringing on third-party patents. It's mandatory for EIC Accelerator grant proposals and demonstrates that you've considered potential IP risks, which is crucial for DeepTech startups.

Since an FTO analysis is mandatory, you'll need to either hire a patent law firm to conduct one or prepare a general analysis in-house. While hiring a firm is the safest option, an in-house analysis is acceptable if you plan to contract a firm later during the project.

The number of patents to discuss depends on the complexity of your technology and the depth of your search. However, a comprehensive FTO analysis should discuss at least 15 patents, especially for complex and sophisticated technologies.

While software patents can be more challenging to obtain in Europe, an FTO analysis is still important. It helps identify potential IP risks and demonstrates due diligence, even if the risk is perceived as lower than for other technologies.

Your FTO should include a company identity section, a statement and disclaimer explaining the analysis's scope, a methodology section detailing your search process, a list of keywords used, a selection and discussion of relevant third-party patents, and an appendix with supporting IP information.

The EIC equity funding process is complex, involving a crowd-in factor, tranches, negotiations, and policy goals. Requesting less equity can sometimes be advantageous, and the EIC should be able to adjust the amount to fit their mission and budget.

Allowing the Jury to adjust the equity amount provides flexibility to account for factors like the crowd-in factor, co-investor requirements, EU interests, and the company's fundraising progress. This prevents companies from being rejected solely due to an unfitting equity request.

Aim for around 8 risks (4 technical, 4 commercial) that are somewhat vague but accompanied by strong mitigation strategies. Avoid focusing on real-world or market risks, as these can lead to negative evaluations. The key is to acknowledge risks without making them seem insurmountable.

While transparency is generally good, being overly transparent about significant risks, especially for groundbreaking DeepTech ventures, can decrease your chances of reaching the interview stage. Focus on presenting risks in a way that demonstrates you've considered them and have effective mitigation strategies in place.

Most companies do aim for the maximum grant amount, and the jury typically doesn't change this amount. However, it's important to carefully consider your actual needs and how the grant will be used to achieve your project goals.

Many applicants are confused about equity requests, focusing more on the grant component. If you're unsure, it's best to seek advice and understand the implications of different equity amounts. The EIC Jury should ideally be able to adjust the amount if needed.

The EIC could adopt a system similar to Airbnb's review process. Step 2 evaluators, who often possess domain expertise, could submit 1-2 questions for the Jury to ask during interviews. This allows the Jury to assess the applicant's understanding and response to technical inquiries.

It's best to avoid emphasizing real-world risks or market risks, such as product-market fit issues. These types of risks can be easily criticized by evaluators. Instead, focus on more general risks with clear mitigation plans.

Domain expertise helps evaluators to assess the technical feasibility and market potential of a project more accurately. Step 2 evaluators often have this expertise, and incorporating their input through targeted questions can improve the overall evaluation quality.

The EIC Jury should focus on circumstances surrounding the risk, such as whether the company has a clear capital need, if there's intense competition, if Europe lacks the necessary ecosystem, or if the value chain is not yet established. These are realistic challenges that the EIC should consider more than just the abstract idea of 'high risk'.

It's recommended to use a European company running an open-source Large Language Model (LLM) locally on EU servers. This ensures data security and avoids sending sensitive Intellectual Property (IP) outside of Europe. A modular setup with separate backend/UI and instructions allows for flexibility and control.

AI enables instant and repeated evaluations, allowing for refinement of the filtering process. It also provides a perfect record of evaluation decisions, enabling continuous improvement based on past successes and failures. Ultimately, the goal is to increase the interview success rate and the quality of winning projects.

AI can be used to automatically send questions to applicants if information is missing or unclear, facilitating a real-time information exchange. The application template could also be replaced with a shorter cheat sheet followed by AI-driven Q&A to streamline the process.

It's crucial to start early to allow for experimentation and development. Many applicants and consultancies are already using AI, and proposal writing is likely to become fully automated soon. Being prepared with a well-developed AI evaluation system will be essential as submission rates increase.

AI should only be fully integrated when it consistently demonstrates higher quality evaluations than human evaluators. Early testing can involve inviting top AI-selected projects that were initially rejected by human evaluators to see if the jury approves.

Beyond evaluation, AI can be used to connect to patent databases, perform web searches for market and competitor analysis, and even replace the traditional proposal template with a more streamlined AI-driven Q&A format.

By saving the core logic and evaluation instructions as plain text, the EIC can easily modify and update the system. This approach allows them to retain control over the evaluation criteria while using external contractors for integration and version control.

Yes, a hybrid approach is possible. AI could generate a dashboard highlighting the pros and cons of each project, acting as the initial evaluators. Human evaluators could then use this information to make the final decision, improving efficiency and consistency.

The interview success rate is the most important KPI. It indicates how well the initial evaluation steps filter out unsuitable projects. A higher interview success rate signifies a more effective evaluation process overall.

By streamlining the application process and potentially replacing it with AI-driven Q&A, the EIC can remove the learning barrier for applicants. This allows the EIC to select from a broader range of European companies, not just those familiar with the EIC's requirements.

The EIC Accelerator should consider prioritizing pure grants and the Grant+ model over equity investments. This approach could be more effective at crowding in private investors, supporting more DeepTech companies, and achieving the program's core mission of bridging the valley of death.

Grant+ acts as a 'VC dowry,' providing a non-dilutive grant to the startup's next funding round. This effectively offers a discount to outside investors, making the company more attractive without requiring them to share equity with the EIC.

By shifting from large equity investments to smaller top-up grants, the EIC Accelerator could fund a significantly larger number of companies. The remaining budget could be used to provide more initial grants, expanding the reach of the program.