What Is Rubric Injection? — How OpenSeed Applies the Evaluation Criteria Written in a Program Announcement to Its Review Process

#What Is a Rubric?

A rubric is the scoring guide mentors use when evaluating applications. It's not a simple good/bad judgment — it's a structured document that specifies how many points are allocated to each evaluation category and what level of quality earns a high score. Korea's Pre-Startup Package, Early-Stage Startup Package, and TIPS programs each publish these criteria, either within the official program announcement or in separate evaluation guidelines.

The problem is that most founders don't cross-reference the scoring breakdown carefully when reading through the program announcement. If "Problem Recognition and Solution" is worth 25 points and "Commercialization Capability" is worth 15 points, you should be allocating significantly more content and evidence to the former. Yet a large share of submitted plans pour the most space into whatever the founder feels most confident about — regardless of how the points are actually distributed.

Knowing the rubric doesn't automatically guarantee a high score. But without it, you can't even diagnose which sections are falling short. That's the starting point.

#How OpenSeed's Rubric Injection Works

Rubric injection means directly incorporating the evaluation criteria, eligibility requirements, and point weightings written in a specific announcement into the review analysis. When a founder selects the announcement they're applying to on the Korean or Japanese home page, OpenSeed pulls in the body of that announcement. From that point on, when up to 17 of OpenSeed's evaluation roles examine their own areas, they diagnose against the evaluation criteria and eligibility requirements written in the announcement.

This goes well beyond checking whether certain topics are mentioned. In the Pre-Startup Package, for example, the "Innovation and Differentiation of the Startup Idea" category carries a substantial share of the total points. With the announcement's criteria in play, each mentor is asking a specific question: "Is this plan concrete and verifiable enough in its differentiation that a real review panel would rate it highly on this criterion?" Generic feedback without announcement criteria never even asks that question.

This is what sets rubric injection apart from a simple checklist comparison. It doesn't just check whether a given category is addressed — it judges whether the writing is convincing enough for a review panel to rate that category highly. Eligibility requirements, however, are checked one by one, and any clear mismatch is designed to be flagged first.

#Pre-Startup, Early-Stage, and TIPS — How the Scoring Structures Differ

These three programs may sound similar, but their evaluation structures are meaningfully different. That's exactly why so many founders make the mistake of submitting the same plan to all three without adjustments. Placing the rubrics side by side makes those differences immediately clear.

Because the Pre-Startup Package targets founders who haven't yet incorporated, track record and team history matter far less than the clarity of the problem definition and the logic of the proposed solution. TIPS, on the other hand, requires prior private-sector investment from an operating company, so its scoring is concentrated on technological innovation and global market potential. A plan that passes the Pre-Startup Package can fail TIPS for exactly this reason — the point distributions are that different.

OpenSeed uses the content of whichever announcement the founder has selected as its standard. If you pick a Pre-Startup Package announcement, the review won't be run against TIPS criteria.

#The Real Advantage for Founders — Compared to Feedback Without a Rubric

The practical value of rubric injection lies in how it changes the direction of feedback. Feedback without a rubric typically sounds like this: "The market size justification is insufficient." Feedback that reflects the announcement's criteria sounds like this (example): "This category carries relatively high weight among the announcement's evaluation items. Specifically, your current write-up is missing a domestic market size figure with a cited source and a clear explanation of how your target segment was sized."

That difference is decisive because it tells founders immediately where to spend their time. The strategy is to address weaknesses in high-point categories first — especially when revision time is limited.

1. Confirm the scoring criteria: Start by understanding the evaluation categories and point weightings for the program you're applying to.
2. Diagnose each category: Assess how well your current plan addresses each scored section.
3. Prioritize high-weight categories: Strengthen specificity, supporting evidence, and data in the categories that carry the most points.
4. Trim over-investment in low-weight categories: If you've given disproportionate space to sections with fewer points, pull back.
5. Check changes yourself: Compare each revised passage line by line against the announcement's evaluation items and confirm that every item now has its evidence and figures.
6. Use the rubric for presentation prep: Once you've cleared the document review, use the scoring structure to prioritize which topics to prepare most thoroughly for your pitch.

This workflow is very difficult to execute without a rubric. Without knowing the point distribution, there's no principled way to set revision priorities.

#Frequently Asked Questions

Here are the questions founders ask most often about rubric injection.

• Q: Which announcements can I pick? — You can pick any open K-Startup announcement that appears in the announcement selector on the Korean or Japanese home page (the selector is not shown on the English home page). The list reflects announcements that were open at retrieval time, and announcements past their deadline don't appear.
• Q: What happens when an announcement changes? — The analysis uses the announcement content registered in OpenSeed, so it may have been revised in the meantime. Check the official announcement yourself before submitting.
• Q: Does the AI finalize an evaluation the way a real reviewer would, using the announcement's point weights? — No. It points out what is missing against the announcement's evaluation criteria and eligibility requirements, but it doesn't issue or predict a pass or fail outcome.
• Q: Can it be used for programs other than these three? — Yes, for any open announcement in the selector list. The analysis also runs if you don't pick an announcement.
• Q: Does an announcement-based review guarantee acceptance? — No. It's a diagnostic tool that identifies gaps relative to the scoring criteria. Final acceptance decisions rest with the actual review panel.
• Q: Can I check my revised plan again in the same report? — There is no way to re-check it in the same report. Check the revised document yourself against the announcement's evaluation items, and apply the report's 'what to do first' items before you submit.

#How to Get Started with Rubric-Injected Review

The process is straightforward. On the Korean or Japanese home page, upload your business plan and pick the announcement you're applying to with the announcement button (the picker is not shown on the English home page). OpenSeed loads that announcement's evaluation criteria and eligibility requirements and diagnoses against them. The report includes key issues, what to do first, and feedback from each mentor, and any item that clearly doesn't match the announcement's eligibility requirements is designed to be flagged first.

The closer you are to a submission deadline, the more valuable rubric-based diagnostics become. The less revision time you have, the more critical it is to know exactly where to focus. Vague, blanket revisions across the whole document won't cut it when time is short. The realistic strategy is to address the weakest points in the highest-weight categories first — precisely and efficiently.

This is especially true for founders who have already been rejected once. When you go back to review a plan after a rejection, it's hard to see what went wrong without a rubric. With one, it becomes considerably clearer which categories likely cost you points.