Frequently Asked Questions

Verify that each evaluation criterion is explicitly addressed with evidence, confirm budgets and timelines match project assumptions, and ensure citations, tables, and figures are present and accurate. Check for originality and avoid generic phrasing that sounds AI-generated by adding specific project anecdotes or pilot data. Finally, confirm compliance with submission rules and required attachments.

That depends on the draft quality: many AI outputs are beyond salvage and require starting over, while others can be improved with targeted edits. Evaluate whether the AI text addresses core evaluation criteria and contains specific details; if not, rebuild using better prompts or a tool that incorporates your data. Reserve editing for drafts that already include solid structure, references, and tailored content.

Start now rather than wait: the landscape is changing rapidly and early applicants can learn how to integrate AI effectively before it becomes chaotic. Build clean project datasets, experiment with purpose-built tools, and iterate on pitch materials so you’re ready when processes evolve. Early preparation also gives you a competitive edge as agencies shift to new evaluation formats.

Start with a structured brief mapping your project to each evaluation criterion, include key facts and references, and request specific sections (methods, impact, budget) in the prompt. Ask the AI to produce formatted elements: tables, charts, hyperlinks, and footnotes. Finally, run a human review focusing on missing evidence, clarity, and compliance before submission.

ChatEIC is built to produce strong drafts by default: it attaches research, includes hyperlinks, formatting, financial projections, and templates tailored to grant sections. It automates common narrative components so you can focus on innovation, customers, and technical details. The result is a more complete draft that typically needs only a final check rather than extensive rewrites.

Generic AI tools often produce short bullets, vague sentences, and miss evaluation criteria because they lack project-specific data and grant-formatting knowledge. They also omit helpful elements like references, tables, and hyperlinks. To avoid sloppiness, augment AI drafts with project data, explicit prompts about evaluation criteria, and a requirements checklist before generation.

Not typically — general chatbots lack domain-specific templates, built-in citation handling, and financial projection features used in competitive grants. They can draft text but often omit essential evidence, formatting, and evaluation alignment. Choose specialized grant-writing tools or augment general chatbots with project datasets and strict templates to get usable outputs.

Agencies will eventually adapt their processes; one predicted change is shifting from essay-based proposals to data uploads processed by AI, with human evaluation focusing on pitch decks, videos, and interviews. In the near term, there will be a mismatch between how applicants generate proposals and how reviewers assess them. Prepare by keeping structured, validated project data and being ready to present concise pitches or demos.

Use AI tools designed for grant writing that ingest your data, include templates, and produce formatted drafts with references and financials. Provide clear, detailed inputs (team CVs, milestones, budgets) and require the tool to output hyperlinks, tables, and citations. Limit manual rework by choosing systems that fill standard sections automatically and let you focus on high-value technical and impact content.

High submission counts from Germany, Italy and the Netherlands reflect active ecosystems and awareness of the call in those countries. However, success is driven by fit to topic and proposal quality, not country of origin. If you’re from a less-represented country, focus on strong technical rationale and partnership quality to compete effectively.

Start by ensuring strict topic fit and tailoring the proposal to evaluation criteria rather than using generic templates. Prepare a convincing business and impact case for Stage 2, and rehearse interview scripts and interviewee selection well ahead of deadlines. Use helpdesks and NCPs for clarifications, and document compliance with call rules to avoid administrative rejections.

For the March Step 2 deadline, applicants were asked to submit interviewee data used for mandatory pre-assessments. This means evaluators now expect detailed information about the people who would attend interviews (roles, expertise, and availability). It’s an added element that teams must prepare carefully because it affects interview readiness and assessment.

If you already submitted and are unsure of eligibility, prepare a strong Stage 1 narrative that explicitly maps your project to the call’s topic requirements. For future calls, contact National Contact Points or the programme helpdesk before applying to confirm eligibility. Keep documentation showing how your project meets topic criteria in case evaluators request clarifications.

Proposals requested a total of €130.7 million. Stage 1 has a fixed budget of €6 million and awards a €300,000 grant per selected project, so approximately 20 projects can be funded. That yields an estimated Stage 1 success rate of roughly 2.8%.

Provide concise CVs, clear role descriptions, and evidence of commitment (availability and decision-making authority) for proposed interviewees. Run mock interviews and align answers across technical, business and impact topics to present a consistent team story. Ensure interviewees know the application inside out and can address weaknesses transparently.

The AIC call received 709 proposals across two topics: 425 for Accelerating Physical AI and 284 for Translating Disruptive NAMs. Applicants came from 39 countries, with Germany (84 proposals), Italy (71) and the Netherlands (62) most represented. Of the 709 submissions, 490 were companies (69%), 124 were from higher education (17%) and 89 were research organisations (13%).

Companies made up the majority of applicants (69%), followed by higher education and research organisations. This dominance by companies suggests strong commercial competition and potentially more experience in proposal preparation. If you’re an academic or research partner, emphasise clear market potential and industrial fit to stand out.

Yes—because the call covered only two focused technology segments, many proposals are likely to be out of scope. Some applicants may have submitted last-minute or without fully checking topic eligibility. It’s important to read the call text carefully and justify how your project fits the specific topic to avoid scope rejection.

AIC results are expected in May 2026. The main benefit of passing Stage 1 is access to Stage 2, which offers much larger support—around €2.5 million per project. Stage 2 is where the programme’s major funding impact and scaling opportunities are realised.

It’s possible processes will change, so don't rely on permanence of any single step. Instead, treat current requirements as real and invest effort in doing them well. Building a repeatable prep routine benefits you regardless of process changes and improves your overall pitching and documentation skills.

Assume your evaluator may not be an expert in your exact field and focus on clarity and impact. Prepare simple explanations of technology, market, and business model, and include a one-page summary so the evaluator can quickly grasp essentials. Use analogies, clear metrics, and visuals in your pitch deck to bridge domain gaps during the 60 minutes.

Random selection can introduce variability in domain expertise, which may affect depth of technical feedback. To mitigate this, ensure your materials are self-explanatory and emphasize practical milestones, risk mitigation, and commercial potential. Also prepare concise answers to common technical, IP, and go-to-market questions to steer the conversation constructively.

Finalize a one-page summary, the pitch deck, and a short FAQ with answers to likely questions on IP, milestones, team, and finances. Do at least two mock 60-minute interviews with different people playing the role of evaluator, and time your answers to stay concise. Prepare a list of supportive documents to share quickly if requested (e.g., FTO summary, key metrics, and roadmap).

Use consultants selectively for high-leverage tasks like refining the pitch, reviewing the business plan, and mock interviews rather than end-to-end writing. Train an internal team member to manage proposal structure and use templates to avoid repeated consultancy fees. Leverage AI tools to draft and iterate documents before hiring experts for final polish.

Use AI for research, drafting annexes, formatting, and generating concise summaries, but always validate facts and refine tone yourself or with an expert. Create explicit prompts and provide your own data so AI outputs are accurate and aligned with your strategy. Reserve human review for technical claims, IP descriptions, and high-stakes financials to avoid errors.

Step 2 pre-assessments are 60-minute remote interviews where each applicant speaks with a randomly assigned evaluator. They are used as an additional evaluation layer before the final steps and simulate the interaction you might have with remote evaluators or jury members. Treat them as formal assessments: prepare a concise pitch, be ready for questions on impact and feasibility, and have key documents and numbers at hand.

Adoption is likely gradual: evaluation bodies are cautious but under pressure to scale and cut costs, so AI integration is probable in the medium term. In the meantime, applicants should assume human evaluators and use AI as an internal productivity tool. Demonstrating that you used structured, evidence-backed data (which AI can help produce) will make your case stronger.

Annexes like freedom-to-operate (FTO) searches, initial literature reviews, executive summaries, and draft investor-style business cases are well-suited to AI assistance. Use AI to generate multiple iterations quickly, then have subject-matter experts verify results and add nuance. This approach saves time while maintaining quality for core technical and legal sections.

Treat AI drafts as first passes: fact-check everything, inject your unique insights, and ensure arguments are evidence-based and tailored to the EIC’s evaluation criteria. Run mock reviews with colleagues or trusted advisors, and refine clarity, metrics, and feasibility before submission. A polished, human-reviewed AI-assisted proposal is far more effective than an unchecked automated output.

Although originally slated for late February or early March 2026, the blog suggests additional weeks of delay due to evaluation complexities; results were expected sometime in the coming week(s) from the post date. Keep monitoring official EIC channels for the formal announcement.

You can sign in with Google or email to streamline access and synchronize saved searches or alerts. Signing in makes tracking calls, receiving notifications, and using personalized features easier.

Scoring relies on written materials and indirect criteria, which makes ranking diverse proposals harder and can lead to repeated review rounds. Without interviews, reviewers must infer program fit from proposals alone, increasing the likelihood of iteration before final decisions.

Subsdy upgraded its grant content to richer overviews that include program documents, templates, and more contextual data for each call when available. This helps applicants find relevant materials faster and better prepare proposals.

No—22 pages is within the expected scope for this pre-accelerator call since it intentionally avoided complex annexes. The issue is not length but the evaluation tension between what the program asks for (higher TRL) and what the modest budget realistically supports.

Moving from TRL4 to TRL6 often requires expensive development, validation, or clinical preparations—especially in MedTech—so €500K tends to cover planning and limited development only. Evaluators must judge proposals against this mismatch, which complicates fair scoring.

The program is new and evaluators are likely working through tricky assessment issues—like comparing diverse proposals without an interview stage and reconciling conflicting requirements (e.g., TRL expectations versus budget). Those procedural and scoring complexities can cause multi-iteration reviews, which explains the delay. The post indicates results are expected within the coming weeks.

The call targets lower-investment regions where lower development costs encourage digital startups, so you'll see more software- and service-oriented applicants. That geographic focus influences the overall mix and could skew evaluations toward lighter-weight projects.

Smaller grants strongly favor early-stage digital or software companies that need funding to build an MVP. DeepTech projects with heavy hardware, regulatory, or clinical work are less likely to be feasible within the budget and therefore may be disadvantaged.

Two additions are noted: EIC Transition (invite-only, up to €2.5M for TRL4→TRL6) aimed at follow-on development, and EIC Scaling Club 2.0 aimed at accelerators for hands-on support. DeepTech teams approaching TRL6 and accelerator operators should follow these calls and subscribe to Subsdy or the newsletter for updates.

You can set your profile to private to turn off profiles and messaging functions, and you can change your claimed handle at any time. Use the privacy toggle in your account settings to disable visibility and communication features.

Agencies should shorten and standardize proposals to essential data fields, limit prose, and impose strict data limits. They should also redesign evaluation steps to rely on structured data and AI-assisted summaries, reducing repetitive tasks for human reviewers.

AI can generate higher volumes of polished, data-rich proposals and perform rapid deep research, which overwhelms manual evaluation workflows. That forces agencies to rethink formats, verification, and scalability because human reviewers can’t keep pace with AI-augmented submissions.

Create a structured data room with mandatory sections (technology, IP, financials, team, market, impact) and hard content limits, then let AI analyze that dataset. Humans would then review AI assessments alongside a short pitch deck and video, greatly reducing time per application and improving consistency.

Fewer evaluators with better tools can evaluate more consistently and faster, while adding many reviewers or manual consultancy steps increases coordination overhead and costs. The goal is to streamline roles and leverage AI for preliminary assessments so humans focus on judgment calls.

The EIC has acknowledged a backlog in evaluating short proposals, which has slowed Step 1 decisions. Increased application volume and new challenges from AI-related proposals have strained existing processes, causing longer wait times than usual.

According to the post, applicants who submitted Step 1 should have received their evaluation results on the noted date. If you haven't received a result, check your inbox and spam folder and contact EIC support with your submission ID for a status update.

Subscribers can create profiles on up to 120 grants, view a grant grid showing recent additions, access a two-month deadline calendar filtered by grants they follow, and see a list of grants where other users are searching for partners.

Free users can now create up to five profiles on any grant page to signal interest and find partners, create reusable profile templates, message other users searching the same grant, and claim a personal handle (subsdy.com/u/handle) to display their grant profiles.

The document section lets grant owners and users attach links to useful materials like templates, work programs, and guidelines. Use it to upload or reference standardized documents applicants need, making it easier for partners to align proposals and for evaluators to find consistent information.

The post highlights that many future technologies (AI, automated manufacturing) require abundant, low-cost, reliable energy, which nuclear power can provide. It notes hyperscalers (Microsoft, Google, Amazon, Meta, Oracle) investing in or partnering for nuclear to secure cheap, reliable power for data centers and AI workloads. The point is that investment in late-stage tech must be paired with energy and policy considerations to be effective.