SB StartupBasket
All ideas
76 /100 GO Medium complexity

ShokeiPrep — handover workbench for retiring Japanese SME owners

Turns a retiring Japanese owner's shoebox of records into the document set a buyer's due diligence actually demands.

ProfessionalServices SaaS Japan SMBAI-agentMultilingualSolo-builderSuccessionDocument-AI
— views
Evaluation Scores
76/100

GO

Overall Score

17
Problem
12
Demand
11
Build
11
Distrib.
12
Revenue
7
Time
6
Defense

ShokeiPrep

1. One-liner

Turns a retiring Japanese owner’s shoebox of records into the document set a buyer’s due diligence actually demands.

2. Trend signal — why now?

Three things moved at once, and they moved toward the seller, which is the side nobody has ever built software for.

The seller side just got government money. In the 15th call of the 事業承継・M&A補助金, announced 22 May 2026, METI’s SME Agency created a brand-new category: 小規模売り手支援類型 — “small-scale seller support.” Subsidy rate 2/3, cap ¥4.5M, and the eligible-expense list explicitly names due diligence, contract drafting, valuation and expert fees. Eligibility is tight and tells you exactly who the state is worried about: ≤20 employees in manufacturing, ≤5 in retail and services. The application window ran 19 June – 24 July 2026. For the first time the subsidy is pointed at the person selling, not the person buying or the intermediary collecting a fee.

The pain is documented and brutal. A selling owner’s first-hand account of his own DD describes the finance-and-tax question set alone as “財務・税務に関する項目だけで80以上の質問と20以上の資料提出依頼が記載されていました” — over 80 questions and 20+ document requests, for one discipline. Then the buyer’s in-house accounting team turned up late with “約30件の質問と15件程度の資料提出依頼” on top. His most telling line is that some of the requested materials simply did not exist and had to be created for the DD. Practitioner guides put the seller’s document scope at six categories including ten years of board minutes, and the elapsed window at 2 weeks to 3 months, run “通常業務と並行しながら” — alongside actually running the company. One firm calls it the 最大の山場, the biggest climb of the whole deal.

And the tooling that shipped in 2026 went to the other side of the table. Herix’s Aidiligence — AI that ingests documents, investigates legal and financial risk and drafts reports — is built, in its own announcement, for “buyers, M&A intermediary IM creation, and buyer screening.” TRANBI added an online DD support tool in 2026, but it is a data room: it shares files that already exist. Nobody sells the seller a tool that produces the file.

Provenance:

3. The opportunity

Japanese SME M&A has an asymmetry that nobody has monetised. The buyer pays for due diligence — that is the standard, and every guide says so plainly: DD費用は全て買い手が負担. The buyer therefore gets professionals, and now gets AI. The seller pays for nothing and gets nothing, yet the seller does all the work: finding the documents, writing the ones that don’t exist, and sitting for the hearings.

The incumbents are all pointed the wrong way. Large intermediaries set minimum success fees of ¥5M–¥25M, which means a company doing ¥80M of revenue is not worth their time and gets turned away. Those owners land on TRANBI and BATONZ, which are matching platforms — they list you, they introduce you, and then they hand you a document request list and wish you luck. Aidiligence reads documents for the buyer. Nobody stands next to the seller.

The failure mode this creates is well documented and expensive: データルームに必要な資料が揃わず「後で出します」と言いながら実質的に出てこない — the data room never fills, “we’ll send it later” turns into never, and the buyer’s trust collapses. Deals die or get repriced. Separately, unpaid overtime is the single most common off-balance-sheet liability in Japanese SME M&A, invisible to financial DD, and when labour DD finds it the purchase price comes down.

So the pitch is not “organise your files.” It is: find what a buyer will ask for, tell you which of it doesn’t exist yet, build it, and flag the two or three things that will cut your price before the buyer finds them. A 10% price protection on an ¥80M deal is ¥8M. Against that, a few hundred thousand yen of software is rounding error — and two-thirds of it may be subsidised.

4. Target market

  • Primary customer: The 社長 of an owner-managed Japanese company, 5–30 employees, ¥50M–¥800M annual revenue, aged 65+, no successor, who has decided to sell and is either self-serving on TRANBI/BATONZ or working with a small regional advisor. Often the wife or a single 総務 person is the entire back office. Manufacturing, construction, transport, clinics, food processing, regional services.
  • Why they buy: In their words, from the practitioner and owner sources: the DD is the 最大の山場; the burden of preparing materials while running the business is 大きな負担; the requested documents sometimes 存在しない and must be written from scratch. They are not afraid of the buyer. They are afraid of looking disorganised, of the price coming down, and of the deal quietly dying because they could not produce a 2019 board minute.
  • Rough TAM reasoning: Roughly 1.27 million Japanese SME owners aged 70+ with no successor. TDB’s 2025 survey puts the national 後継者不在率 at 50.1%. Japan recorded 5,115 domestic M&A deals in 2025, a second consecutive record. I do not need the 1.27M — I need the slice that actually transacts or seriously tries: tens of thousands of seller-side processes a year, plus the far larger pool of owners in the 1–3 years of preparation before listing. At ¥40K/month for six months, a few thousand paying owners is a real business.
  • Why now for them: The subsidy category aimed at them was created in May 2026. The buyer across the table now has AI reading their documents. And the owner is 68, not 58 — the window is closing on its own schedule.

5. Product sketch (MVP)

  • Buyer question set, pre-loaded. The standard Japanese SME DD request list across the six disciplines — 財務・税務, 法務, 労務, 事業, 許認可, IT — rendered as a live checklist, not a PDF. Roughly 130 items, matching what real sellers report receiving.
  • Photograph the archive. The owner or their 総務 staffer shoots the filing cabinet, the ledger binders, the 決算書 stack, the 就業規則 folder. The system reads scanned, sealed and handwritten Japanese records and files each page against the checklist item it answers.
  • The “does not exist” list. The output the owner actually needs: not what they have, but what the buyer will ask for that isn’t anywhere in the building. Missing board minutes, missing employment rules, no organisation chart, no customer contract file, no equipment ledger.
  • Draft the missing documents. From what the archive does contain, produce first drafts of the perennial gaps — 組織図, 取引先一覧, 設備台帳, 株主名簿 reconstruction, 議事録 for years where only fragments survive — for the owner and their 顧問税理士 to check and sign.
  • Price-cut early warning. A focused pass over attendance records, 就業規則 and payroll for the unpaid-overtime exposure that is the number-one off-balance-sheet liability in Japanese SME deals, plus expiring 許認可 and un-updated 株主名簿. Quantified, with a range, before the buyer’s 社労士 finds it.
  • Readiness score and Q&A drill. One number the owner can watch move, and the follow-up questions a buyer typically asks after each answer, so the hearing is not the first time they hear them.
  • Subsidy packet. Assembles the documentation for a 小規模売り手支援類型 application, including the two comparative quotes the scheme requires for expenses under ¥5M.
  • Hand-off export. One structured package for the intermediary, the buyer’s data room, or the 顧問税理士.

6. AI angle — what’s load-bearing

Remove the AI and this is a checklist PDF, which already exists free from a dozen intermediaries and helps nobody, because the work was never the list — it was reading twenty years of paper.

Three jobs are AI-only at this price point. Reading the archive: Japanese SME records are scanned, handwritten, and stamped with 印鑑 that historically wrecked OCR accuracy; current vision models handle messy real-world Japanese documents well enough to route pages to checklist items, and that changed recently. Inferring absence: telling an owner what is missing requires understanding what each document is and what a complete set looks like — that is a reasoning task over an unstructured pile, not a lookup. Drafting the gaps: generating a defensible 組織図 or 設備台帳 from invoices, payroll and photographs is exactly the work that made the seller’s own account say the materials had to be created from nothing.

A 社労士 or 税理士 can do all three. They charge ¥300K–¥1M and take weeks, and the small seller was already priced out of professional help — that is why they are on a matching platform in the first place.

7. Localization angle

This is Japan-only and should stay Japan-only, which is the whole point.

The document set is Japanese-statutory: 決算書, 総勘定元帳, 就業規則, 労働者名簿, 株主名簿, 議事録, 許認可. The failure modes are Japanese: unpaid overtime as the dominant simulated liability, shareholder registers that were never updated after a 1990s transfer, 名義株 held by relatives. The buyer’s expectations are set by the 中小M&Aガイドライン. The subsidy is a METI programme with Japanese eligibility thresholds and a Japanese application form. The distribution runs through 事業承継・引継ぎ支援センター in all 47 prefectures, 信用金庫, and regional 税理士.

Price at ¥39,800/month and it lands where Japanese SME SaaS lands — comparable to a mid-tier 会計 or 労務 subscription, and small against a ¥5M minimum intermediary fee. A US-shaped $500/mo product with English document handling is not a competitor here; it is not even a substitute.

The obvious later market is Korea and Taiwan, both facing the same demographic wall with the same unprepared owners. Different documents, same product shape. Not year one.

8. Business model — path to $1M–$5M ARR

  • Pricing: ¥39,800/month (~$270) for the active preparation period, typically 4–8 months. Plus a ¥150,000 one-time 初期整備 fee covering the bulk archive ingest — the heavy compute month. Advisor tier at ¥98,000/month for 税理士 and 社労士 firms running multiple seller clients.
  • ACV: A direct owner runs roughly ¥400K over a 6-month engagement (¥150K setup + ~¥240K subscription). That is the honest number, and the honest problem: it is a finite engagement, because the business gets sold and the customer disappears. Advisor-tier firms are the recurring revenue.
  • Rough math to $1M ARR: ~¥150M. Either ~375 owner engagements a year at ¥400K, or a blend — 200 owner engagements (¥80M) plus 60 advisor firms at ¥98K/month (¥70M). The blend is the realistic path and the advisor half is the durable half.
  • Rough math to $5M ARR: ~¥750M requires ~300 advisor firms and ~1,000 owner engagements annually. That means being the default seller-prep tool inside the 税理士 and 事業承継支援 channel nationally, plus a second market — Korea most likely — carrying maybe a quarter of it. Reachable in year 4 if the advisor channel compounds; not before.
  • Expansion path: Owner engagements are the acquisition wedge and the case-study factory; advisor seats are the annuity. Post-close there is a real second act — the PMI document handover the buyer needs on day one, sold to the buyer with the seller’s blessing. And the un-listed owner two years from sale is a lower-priced 磨き上げ tier: fix the shareholder register and the employment rules now, while it is cheap.

9. Go-to-market wedge — first 100 customers

The owners are old, offline, and reachable only through people they already trust. Do not run ads at them.

  • The 税理士 channel is the whole game. Every one of these companies has a 顧問税理士 who visits monthly, is the single most trusted outside adviser, and is personally embarrassed when DD surfaces a mess on their watch. There are ~80,000 registered 税理士 in Japan, heavily concentrated in small regional practices. Target 500 firms in three prefectures with a specific offer: a free readiness diagnostic on one of their client companies, run by us, delivered as a report they hand over with their name on it. Advisor tier converts from there. Expect 500 approaches → 60 diagnostics → 15 advisor firms, each carrying 2–5 owner engagements.
  • 事業承継・引継ぎ支援センター in all 47 prefectures. Government-funded, mandated to help exactly these owners, chronically under-resourced, and they refer out constantly. These are relationship builds, not sales calls — one prefecture at a time, starting with the highest 後継者不在率 regions (Akita at 73.7% is the extreme). A center that adopts the readiness checklist sends a steady trickle for years.
  • Subsidy-deadline outreach. The 小規模売り手支援類型 runs on published application windows. Owners and their advisers are actively hunting eligible expenses in the weeks before a deadline. Publish the definitive Japanese-language guide to the category — eligibility thresholds, the two-comparative-quotes rule, what counts — and be the tool that assembles the packet. This is the one channel where inbound actually works, because the search intent is sharp and dated.
  • 信用金庫 and regional bank succession desks. They hold the lending relationship, they know which owners are 68 with no successor, and their succession desks are under pressure to show activity. Slower than the 税理士 channel and more bureaucratic; start after the first case studies exist.
  • Case studies, in Japanese, with numbers. “The buyer’s 社労士 would have found ¥12M of unpaid overtime; we found it first and the owner disclosed it on their own terms.” That story, told to 税理士 study groups, is the entire marketing engine. This channel is trust-based and slow to start, then compounds.

10. Build complexity — justification

Medium, and honestly at the upper edge of it. The document pipeline is off-the-shelf — vision models handle Japanese scanned and sealed records adequately now — but the domain content is the real work: encoding a genuine 130-item DD request set across six disciplines, knowing what a complete Japanese SME document set looks like, and drafting 議事録 and 台帳 that a 税理士 will actually sign off on. That is not something a generalist builds from a blog post; it needs a 税理士 or M&A practitioner co-founder or a very well-paid advisor from week one.

Call it 4–5 months to a credible v1 for two people, one of whom has real Japanese SME domain depth. The unpaid-overtime exposure module is the piece I would build second, not first — it is the highest-value output and the easiest to get embarrassingly wrong.

11. Gating checklist

GatePass?Note
Legal in target market✅Document preparation and organisation. Carefully not 税理士業務 or 社労士業務 — outputs are drafts for a licensed professional to review and sign. That boundary is a real design constraint, not a formality.
Ethical — no harm / dark patterns✅Pushes sellers toward earlier and fuller disclosure, which is what the 中小M&Aガイドライン wants. The product’s value is surfacing problems before the buyer does, not hiding them. Any feature that helped conceal a liability would be both unethical and a lawsuit.
Market exists (evidence above)✅5,115 deals in 2025, 50.1% 後継者不在率, a subsidy category created specifically for this buyer in May 2026.
1–5 person team can build this✅Two people, one with domain depth.
Launchable with <$50K / ₹40L✅Document AI inference on a few hundred archives, a Japanese-language site, and travel to three prefectures. The expensive input is domain expertise, not capital.

12. Feasibility score

AxisWeightScoreNotes
Problem intensity2017/20The 最大の山場 of the biggest financial event of the owner’s life, run alongside a full-time job, with a documented price-cut mechanism when it goes badly. Not 18+ only because it is once-in-a-lifetime, not recurring — there is no daily bleed to point at.
Demand evidence1512/15Strong and specific: first-person seller accounts, quantified question counts, a purpose-built subsidy, national deal statistics. Held under 13 because I have practitioner and owner testimony about the pain but no direct evidence of owners paying for this shape of software — it doesn’t exist yet.
Build feasibility1511/15Pipeline is off-the-shelf; the domain content is not. Handwritten and sealed Japanese archive material remains the technical risk. 4–5 months, and it needs licensed-professional review built into the workflow.
Distribution clarity1511/15The 税理士 channel is named, countable and correct, and the subsidy-deadline play is genuinely sharp. Docked because trust-based professional channels in Japan are slow to open, and a foreign or unknown founder will feel that acutely.
Revenue mechanics1512/15¥39,800/mo is well-benchmarked and trivially justified against a ¥5M minimum intermediary fee. The structural flaw is real: the owner customer churns by design when the company sells. Advisor seats fix it, but that is an assumption, not a fact.
Time to first revenue107/10A hand-run diagnostic can be sold within 6–8 weeks. Full product revenue is more like 4–5 months, and the 税理士 relationship cycle adds weeks.
Defensibility106/10The moat is the encoded request set, the accumulated map of what Japanese SME archives actually contain, and 税理士 channel relationships. All soft. A Japanese incumbent — freee, Money Forward, or BATONZ itself — could build this if they decided the seller was worth serving. So far none of them has.
Total10076/100

13. Qualitative modifiers

Founder-fit tags

technical-heavy · domain-expertise-required

Native-level Japanese is non-negotiable, and so is a credentialed domain partner. A foreign solo builder should not attempt this one.

Key assumptions to validate (3–5)

  1. Assumption: Owners of ¥50M–¥800M companies will pay ¥39,800/month during preparation rather than push the work onto their 顧問税理士 for free. How to test: 20 structured interviews with owners currently listed on TRANBI/BATONZ, presenting the readiness report and the price. Watch for the “my 税理士 does that” reflex and whether it survives being shown the 130-item list.
  2. Assumption: 税理士 firms will treat this as a client-service upgrade rather than a threat to billable work. How to test: Pitch 30 regional firms. The tell is whether they want the advisor tier or ask to white-label it — and whether anyone gets defensive about the drafting features.
  3. Assumption: The document pipeline actually handles real Japanese SME archives — 1990s handwritten minutes, sealed contracts, faded thermal-paper ledgers — at a quality that survives 税理士 review. How to test: Get 5 real archives under NDA. Measure item-routing accuracy and, more importantly, whether the “missing documents” list is right. A false “you have this” is worse than a false “you’re missing this.”
  4. Assumption: The unpaid-overtime exposure estimate can be made accurate enough to be useful without crossing into 社労士 practice. How to test: Run it against 5 cases where a real 労務DD was completed and compare the ranges.

Risk flags

  1. Regulatory / professional-practice risk: The line between document preparation and licensed 税理士業務 / 社労士業務 is the product’s central legal constraint. Get a written opinion before launch, and design the review-and-sign workflow around it rather than bolting it on.
  2. Customer churn by design: Every successful engagement ends with the customer selling their company and leaving. Without the advisor tier this is a treadmill, and the treadmill speeds up as you grow.
  3. Incumbent wake-up: BATONZ is listed and well capitalised, and freee and Money Forward already own the SME accounting relationship. Any of them could ship a seller-prep module. The counter is that all three have spent years pointing at the buyer and the intermediary — but that is a habit, not a barrier.
  4. Market timing, in both directions: The demographic wave is real but the sale is a one-time decision an owner can defer for years. Meanwhile the underlying urgency is finite — the 1.27M cohort works through this over a decade, not forever.
  5. Foreign-founder friction: This market runs on introductions, prefectural relationships and 信用金庫 desks. Without a Japanese co-founder the distribution plan in section 9 does not execute.

14. Structured verdict

Score:                  76/100
Verdict:                GO
Confidence:             Medium
Best-fit builder:       Japanese technical founder paired with a 税理士 or SME M&A practitioner;
                        native Japanese and regional relationships are prerequisites, not advantages
Time to revenue:        8 weeks for hand-run diagnostics; 4–5 months for product revenue
Capital to launch:      ¥4–6M ($27–40K), majority of it the domain co-founder's runway
Top 3 assumptions to validate first:
  1. Owners pay ¥39,800/mo rather than defaulting to their 顧問税理士 — 20 interviews with
     currently-listed TRANBI/BATONZ sellers
  2. 税理士 firms adopt the advisor tier rather than seeing it as billable-work erosion —
     30 regional firm pitches
  3. Document pipeline reads real 1990s-era Japanese SME archives accurately enough that the
     "missing" list is trustworthy — 5 real archives under NDA
Kill criteria:
  - Abandon if fewer than 3 of 20 interviewed owners will pay ¥39,800/mo after seeing the
    full 130-item request list
  - Abandon if fewer than 2 of 30 pitched 税理士 firms take the advisor tier at ¥98,000/mo
  - Abandon if the missing-document list is wrong on more than 20% of items across 5 real
    archives — an owner who trusts a false "complete" and gets caught in DD is worse off
    than one who never used the product
  - Abandon if BATONZ, freee or Money Forward ships a seller-side readiness module before v1

15. Next step — 1-week validation sprint

  • Day 1–2: Build the 130-item request set for real, from the published DD checklists of five intermediaries plus the 中小M&Aガイドライン. Not a sketch — the actual artefact, in Japanese, that you will put in front of an owner. This alone is worth having.
  • Day 3–4: Get 5 real archives under NDA — through one friendly 税理士 or a retiring owner in the network — and run the pipeline. Measure two things: routing accuracy per item, and the false-complete rate on the missing list. The second number is the one that matters.
  • Day 5: Show the readiness report and the ¥39,800 price to 20 owners currently listed on TRANBI or BATONZ, and pitch the advisor tier to 10 regional 税理士 firms.

Falsifiable outcome: ≥3 of 20 owners commit to ¥39,800/mo, ≥1 of 10 税理士 firms commits to the advisor tier, and the false-complete rate on the missing-document list is ≤20% across the 5 archives. Miss any of the three and this goes back on the shelf — the owner-side willingness to pay is the axis I am least sure of, and the one that kills it fastest.

Interested in a detailed proposal?

Get a deep-dive with market research, competitive analysis, and implementation roadmap.

Contact us

info@startupbasket.ai