The Business Plan Mistake Log Series — 4 Point-Losing Patterns AI Review Keeps Finding

#Why look at the 'pattern' first

More often than not, the reason a business plan gets rejected has less to do with the idea itself than with how the document is written. Reviewers have no way to validate an idea's real-world potential firsthand — which means they can only judge based on the evidence and logical structure written into the plan. A perfectly executable idea gets rejected because of a sloppy document, over and over again.

OpenSeed's AI review is built as a critique-first structure that surfaces weaknesses before praise. That structure happens to create an ideal vantage point for observing which types of problems keep showing up in submitted plans. This Mistake Log series is the organized output of that observation.

Knowing the pattern in advance means you can defend against it. Knowing which items make reviewers suspicious, and which ways of writing something erode trust, means you don't have to repeat the same mistake. This article is the starting line.

#The 4 point-losing pattern families — the full map

The point-losing factors that keep recurring across OpenSeed reviews fall into four broad families. The table below lays out the core issue, representative symptoms, and review impact for each family. Each family gets its own dedicated deep-dive article in this series.

These four families aren't independent of each other. Weak market-size evidence drags down trust in your execution plan's target numbers too. Skip your risk assessment, and it often spills over into poor regulatory awareness as well. Understanding how a problem in one family bleeds into another matters.

#A preview of each family — where to read what

This series runs four in-depth articles in total. Each one covers the real shape of the problem in that family, how reviewers actually react to it, and concrete directions for improvement.

① Weak market-size evidence is the most frequently occurring pattern. Plenty of founders write down TAM, SAM, and SOM figures; few show how those figures were actually derived. It's a recurring pattern to see only a top-down estimate — 'we'll capture X% of the total market' — with no bottom-up estimate (customer unit price × expected customer count) built up alongside it. For the fundamentals of market sizing, see our existing articles 'A Practical Guide to Calculating TAM/SAM/SOM' and 'Bottom-Up Revenue Estimation.'

② Thin risk assessment shows up most clearly in the competitor-analysis section. It's common to see a plan list competitors while only emphasizing its own advantages, never addressing how a competitor might strike back or where the company itself is genuinely weak. Reviewers are checking whether a founder is looking risk in the eye. Give the impression you don't know your own risks, and trust in your team's capability wavers right along with it.

③ Poor awareness of industry regulation shows up repeatedly in regulation-heavy industries like healthcare, finance, food, and mobility. If a plan never mentions the licenses or filings required before launch, reviewers tend to conclude the founder doesn't understand the structure of that industry at all.

④ Unconvincing execution plan shows up concentrated in the milestones and budget sections. Vague timing language like 'launch the service within Q1,' and a funding plan that states only totals for labor, outsourcing, and marketing costs with no supporting rationale, are recurring patterns. For execution-plan fundamentals, see our existing articles 'Business Plan Milestones and Timelines' and 'Funding Requirements Planning.'

#Self-check before you submit — a checklist across all 4 families

Before you submit your plan, running through the checklist below lets you filter out the basic problems in all four families yourself. Being able to answer 'yes' to every item is the minimum bar for defending against point loss.

1. [Market size] Your TAM/SAM/SOM figures have a cited source and a stated derivation.
2. [Market size] You've paired your top-down estimate with a bottom-up estimate (customer unit price × expected customer count).
3. [Market size] You've described the market share you can realistically capture within 3-5 years at a believable level.
4. [Risk] You honestly acknowledge major competitors' strengths, and describe how you'll respond.
5. [Risk] You've named at least two factors that could cause the business to fail or stall.
6. [Risk] You've included a response plan or mitigation strategy for each risk.
7. [Regulation] You've mentioned the licenses, filings, or qualifications required to operate the business.
8. [Regulation] You've named the timeline for obtaining that approval and who (team member or outside expert) owns it.
9. [Execution plan] Your milestones are specified at the monthly level or finer, not lumped into quarterly blocks.
10. [Execution plan] Each funding category (labor, outsourcing, marketing, etc.) has a stated basis for its dollar amount.
11. [Execution plan] Your hiring plan is concrete about role, timing, and compensation level.

If you hit an item where you genuinely aren't sure, that's very likely the exact spot a reviewer will be suspicious of too. Fixing whichever item you got stuck on first is the most efficient place to start.

#How to use this series

You don't need to read the Mistake Log series in order. It's more effective to jump straight to the detailed article for whichever family you couldn't answer 'yes' to on the checklist above. That said, if you're a first-time founder just starting to write your plan, we'd recommend starting with family ① on market size — if you can't define your market, the writing in every other family tends to wobble along with it.

If you already have a draft plan, another path is to run OpenSeed's AI review first. Check which family the report's critiques cluster around, then reference that family's detailed article for your revisions — a more resource-efficient way to work.

#Frequently Asked Questions

Q. What kind of grant programs do the patterns in this Mistake Log series apply to?

A. These patterns were mainly observed around government startup support programs (Korea's pre-startup and early-startup package programs, TIPS, and similar). That said, the four families themselves are universal problems that apply just as much in VC and accelerator review. Specific evaluation criteria vary by program, so it's worth cross-referencing the actual program announcement alongside this series.

Q. How closely does the AI review match an actual reviewer's judgment?

A. AI review centers its evaluation on logical structure and whether evidence is present. Real reviewers weight the same criteria heavily. That said, variables like an individual reviewer's field experience or policy priorities are hard for AI to fully capture. It's best to treat OpenSeed's results as a reference benchmark, used alongside a cross-check against each program's specific evaluation criteria.

Q. Is this series useful for revising a plan that's already been rejected once?

A. Yes. Since most rejection notices don't spell out a specific reason for losing points, re-reading your plan against these four families makes it much easier to spot the problem yourself. If you're stuck on 'where do I even start fixing this, and how much,' getting a family-by-family diagnosis from OpenSeed's AI review first can be a good starting point.