The EIC Accelerator funding (grant and equity, with a blended financing option) by the European Commission (EC) and the European Innovation Council (EIC) awards up to €2.5 million in grant and €10 million in equity financing per project (€12.5 million total). This article provides a perspective on the use of Artificial Intelligence (AI) in grant writing and its impact on startups and Small- and Medium-Sized Enterprises (SMEs), as well as professional writers, freelancers, and consultants. Find ChatEIC here.

The EIC Accelerator Funding and AI Grant Writing

AI has evolved greatly since ChatGPT crashed onto the scene in 2022 and disrupted people’s minds more than it disrupted any industries. But three years later, disruption has unquestionably occurred. Companies like Klarna were able to greatly downsize customer support teams, and Meta’s Marc Zuckerberg said that a large part of their code base is now written by AI.

With every IT company in the world dipping their toes into the AI ocean, many different features and use cases have emerged, with some features being introduced and phased out before companies can even think of relying on them (i.e., OpenAI phases out Assistants API in favor of Responses API).

But within that continuous evolution emerges a new challenge: When building an app, which type of AI interaction is the most useful or appropriate? A chat interface with attachments? An agent with access to files and screens? A live-streamed voice for immediate responses? Deep thinking or web search? Image generation or coding?

Different features give rise to different use cases, but which type is the most effective when it comes to AI grant writing?

AI Grant Writing: Chatbot, Agent or Diverse?

There are many different ways of integrating AI, with chatbots being the first example that showcased its usability. Customer support AI, as spearheaded by Klarna, which acted as OpenAI’s guinea pig, was the obvious disruption target for this mode of interaction, but it was quickly recognized that there were limitations to the chat interface and an inherent ceiling to what could be achieved with it.

Nonetheless, the format has evolved greatly, and features such as DeepResearch (i.e., searching and processing dozens or hundreds of sources) and Thinking (i.e., analyzing queries on multiple levels and questioning response options) have given it more vigor. Other features, such as OpenAI’s canvas, embed more collaboration into the chat interface, while the voice mode offered by many AI companies delivers a more personal user experience.

There are numerous other examples of what the chat format can achieve, but it is clear that its current state of the art is not the end of the line. On the contrary, with more ideas and innovations, the possibilities are endless, and its potential might never be exhausted.

It seems like every week a new AI company, usually from China or the US, enters the global conversation with a new flagship product that seems to be superior to others in that niche. Sesame AI made waves for its conversational AI model (see this video), DeepSeek broke into the scene with a low-cost model that seemed to be on par or even exceed the state of the art, and Manus AI showcased its agentic AI, which could perform complex tasks while accessing the user’s computer.

Where Is Europe In This Discussion?

While Europe is rarely part of this conversation, it does give hope since it shows that innovating in AI does not need to be expensive or, at least, it does not have to be very expensive. With enough ingenuity and grit, it is possible to become a leader in a specific use case, despite the head start of others.

Whether these innovators will be able to hold their position in their use cases is a different question, but it shows that the large market leader cannot be everywhere at all times. Even with their near-unlimited capital, they still have to budget and focus well. This leaves gaps and new innovations to be exploited.

Europe can produce leaders in AI, but it should try to create new things rather than follow (and lag) behind the others.

Innovation Is Always Possible

xAI managed to create large superclusters at such a scale that it gave the AI world a scare since it was deemed impossible to have so many coherent GPUs operate in harmony. And they are still scaling further.

DeepSeek managed to create a highly ranked flagship model based on a fraction of the compute used by leaders such as OpenAI or Anthropic. While they might not have been entirely transparent about the GPUs they actually used due to restrictions on Chinese imports (i.e., potentially bypassed via Singapore), it is true that DeepSeek innovated greatly.

These examples should be an inspiration to Europe since, if they can do it, Europe surely can do it as well.

Form Factors: Which Is Best?

With all these formats at our disposal, choosing the right one for the job is difficult. It might seem like everything is converging into agents as the dominating format due to their comprehensive access, but it might not be the obvious choice in some use cases.

For AI grant writing, as for any other use case, the best format of AI interaction is the one that leads to the user’s goal with maximum efficiency and little friction. For an AI proposal writer, especially in the complex EIC Accelerator grants, this would require the following parameters to be aligned:

User Knowledge: The user must be able to produce a proposal with zero knowledge. Ideally, the user must be able to produce the proposal and not be forced to learn about the EIC in order to complete it, since the point of the AI writer is to align the proposal sections to EIC rules for them. The user does not wish to become a grant writer; they wish to create a proposal.
Output: The user should be able to get to the final product with very little friction. The AI should not ask the user for their opinions since this will be overwhelming. Just like consultants and professional writers provide guidance and a firm direction, the AI must do the same.
Adjustments and Editing: There should be a way for users to edit proposal sections that need updating or to use the AI to make new edits for them, if they are substantial.

With these constraints, it is easier to develop an approach for AI grant writers.

AI Grant Writing: Pragmatic Approaches

The most pragmatic approach that fulfills the above conditions is an AI interface that allows input data (i.e., text and files) alongside a single button that can create an entire proposal (see ChatEIC). That is not only a simple approach, but it is also effective since it requires no user knowledge, generates quality output, and allows the embedding of placeholders as well as a word file generation (i.e., .docx) for later edits. It can likewise allow individual parts of the proposal to be regenerated, if necessary.

This approach only requires simple AI integrations since a chat-like response is sufficient to combine user input with comprehensive and effective prompts to reach a high-quality output. It likewise avoids the conversational part of the AI system (i.e., chatting with AI) since this is not only unnecessary but also makes the output worse (i.e., long conversations dilute instructions and input data).

While ChatEIC 1.0 was a simple GPTs enabled through OpenAI (i.e., a custom chatbot), ChatEIC 2.0 is a stand-alone web app that creates entire EIC Accelerator proposals in under 10 minutes (i.e., from start to Word-file download) while using AI APIs in the backend.

The Next Level: ChatEIC 3.0

While ChatEIC 2.0 is still evolving, it is possible to conceptualize what its future might look like. Agentic AI is an incredible format, and it will surely be as embedded into professional and casual usage in the future as the chatbot interfaces are today. Agents can be viewed as the next level since they do not wait for your response but have full access to act as you, as your agent.

Agents can access your files, use your credit card, create complex errands, and more. Tools such as Cursor AI are using agents for coding and enable the AI to manage an entire code base instead of just being an on-demand chatbot like Google’s Gemini, which has its AI trapped in a sidebar, unable to edit the documents it is positioned in.

It is therefore foreseeable that agentic AI tools will be the future of grant writing. It will be possible to have an AI EIC expert not only draft an entire proposal but also re-evaluate how certain sections could be enhanced to increase cohesiveness and quality.

While this is not possible today due to high context window requirements, domain expertise, and other difficulties, it will be in the future.

The EIC’s Modus Operandi: Cover And Move

One of the reasons why there will not be any big AI innovations, particularly for the EIC, is that the user base is too small. There are just thousands of applicants per year compared to millions or billions for other use cases, where millions and even billions of users are reachable.

A second reason is the EIC’s track record of changing its process. In 2018, the EIC introduced interviews as part of the overall proposal evaluation process. In 2019, it introduced the EIC Accelerator as a rebrand from the SME Instrument Phase 2. In 2021, it introduced the EIC’s AI platform, which was neither AI nor a platform but a collection of form fields with pretty CSS and a clunky back-end (see The EIC’s AI Disaster). In 2023, its AI platform had a spectacular implosion just days before a hard deadline, leaving its applicant base in a scramble.

Since then, things have been peaceful, but it is likely that the next black swan is already brewing behind the scenes. Normally, a black swan event should be a rare occurrence, but the EIC seems to prefer a more frequent appearance since all of the changes generally come with a restructuring of templates and, if written templates are being restructured, applicants and consultants have to adjust everything.

But these changes are hurting applicants the most since, while consultants are in the loop, applicants are entirely confused by contradicting data. Europa.eu pages by the EIC will feature outdated content that ranks high in search engines. This misinforms applicants regarding old Technoology Readiness Levels (TRLs), templates, deadlines, rules, and more.

This is also an obvious obstacle to any AI development since a small user base combined with an uncertain environment does not justify complex integrations or developments.

Conclusion: Robustness Over Fanciness

When it comes to EIC grants, robustness is preferable over fancy features since it is not a forgiving subject to focus on. The EIC can be akin to a ghost ship drifting through the cloudy sea, aimlessly. It introduces changes that seem random and inconsistent (i.e., TRL changes, funding types, template structures, etc.) while its leadership seems entirely divorced from the experience of the majority of applicants, namely the 95%+ that are rejected.

Nonetheless, the EIC has been endowed with a €1 billion budget, provided by taxpayers, and still presents a great opportunity for entrepreneurs all over Europe. For an AI grant writer, the best approach is to simplify the process as much as possible and be versatile enough to ride the waves of change that the EIC will surely stir up every other year.

And with a robust base, an AI grant writer such as ChatEIC will be able to evolve since every disruption will clear redundancies and yield a more focused and targeted tool.

This article was last modified on Apr 2, 2025 @ 09:01

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:

Step 1
- Open now: Apply as soon as possible to be eligible for the next Step 2 submission deadline
Step 2 (closing 17:00 Brussels Time)
- 1^st cut-off 2025: -
- 2^nd cut-off 2025: March 12^th 2025
- 3^rd cut-off 2025: -
- 4^th cut-off 2025: October 1^st 2025
Step 3
- 1^st cut-off 2025: -
- 2^nd cut-off 2025: TBD
- 3^rd cut-off 2025: -
- 4^th cut-off 2025: TBD

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!

Get ChatEIC - The EIC Accelerator Grant Writer here:

by Stephan Segler, PhD
Professional Grant Consultant at Segler Consulting

General information on the EIC Accelerator template, professional grant writing and how to prepare a successful application can be found in the following articles:

A Quick FTO Guide for EIC Accelerator Applicants in a Rush

2023 Budget Allocations for EIC Pathfinder, Transition and Accelerator

Developing the Unique Selling Points (USP) for the EIC Accelerator

Explaining the Resubmission Process for the EIC Accelerator

A Short but Comprehensive Explanation of the EIC Accelerator

EIC Accelerator Success Cases

Deciding Between EIC Pathfinder, Transition and Accelerator

A Winning Candidate for the EIC Accelerator

EIC Accelerator Interview Preparation Process: Scripting the Pitch (Part 1)