For a buyer who wants to "reduce headcount with a system" in the face of a labor shortage, there is only one thing that matters about the SME Labor-Saving Investment Subsidy. If an off-the-shelf general-purpose product (a ticket-vending machine, a food-serving robot, and the like) is enough, the light-procedure "catalog-order type" fits; if you need system construction tailored to your own operations or automation via AI, the "general type" — which places "machinery and system construction costs" as a mandatory expense — fits. Decide this fork first, and every other requirement hangs off it. This article is not about application techniques for the program; it is a neutral explanation to help a buyer decide what to invest in, how much, and toward what effect target. All figures are based on the official application guidelines (from the SMRJ, the Organization for Small & Medium Enterprises and Regional Innovation, Japan), but the subsidy rate, ceilings, and schedule are revised each application round, so always make your final check against the latest version.
1. The bottom line first: choosing between the catalog type and the general type
The Labor-Saving Investment Subsidy is a program that subsidizes part of the cost for SMEs struggling with a labor shortage to introduce "equipment and systems for labor-saving." There are two entry points, and the first thing a buyer must decide is this fork.
| Aspect | Catalog-order type | General type |
|---|---|---|
| What you buy | Choose a general-purpose product already registered in the catalog (IoT, robots, etc.) | Made-to-order equipment and system construction tailored to your own operations |
| Investments it suits | Off-the-shelf products such as ticket-vending machines, food-serving/cleaning robots, and self-checkout machines | Custom builds such as business-system development, AI automation, and DX |
| Procedural weight | Light (the sales vendor supports the application; rolling intake) | Heavy (a business plan and screening; round-based intake) |
| Mandatory expense | Purchase of a catalog product | Machinery and system construction costs (mandatory) |
| Rough subsidy ceiling | Up to the order of 1,000万円 by employee-size bracket | Up to the order of 1億円 by employee-size bracket |
| Rough subsidy rate | 1/2 or less | 1/2 (2/3 for a large pay raise) / 2/3 for small-scale businesses |
Software-centric labor-saving — for example, moving order-taking, inventory, and dispatch management onto a custom B2B SaaS — does not exist in the general-purpose catalog, so as a rule it rides on the general type. Conversely, if you are replacing physical on-site work with an off-the-shelf robot, the catalog type is overwhelmingly faster. It is not that one is "correct"; the first cut is simply whether the target of your investment is in the catalog.
2. The program's purpose: labor shortage × labor-saving × pay raises
This subsidy is not merely capital-investment support. The aim the officials set out is a single through-line: "use labor-saving investment to raise value-added and productivity, and connect that to pay raises." This ties directly to the effect measurement and requirements discussed below.
What buyers easily misunderstand is that it is not "a subsidy for cutting people." What the program measures is labor productivity (value-added ÷ labor input), and investments are valued not only for shrinking the denominator (person-hours) but also for producing more value-added with the same headcount. Labor-saving does not equal layoffs; the program is framed as investment "to redirect a limited headcount toward higher value-added work amid a labor shortage."
For this reason, an application under either type carries a productivity-improvement target, and the general type clearly imposes pay-raise requirements. It is not "install the system and you're done"; unless you are a buyer who can explain "how productivity rises as a result of installing it," you do not fit the program's intent in the first place.
3. The catalog-order type — labor-saving "right now" with off-the-shelf general-purpose products
The catalog-order type is a scheme where you choose and introduce, from a "catalog," a general-purpose product that has been registered and screened in advance. In the official wording, you "select and introduce, from the catalog, general-purpose products such as IoT and robots that are effective at raising value-added and productivity." Its greatest advantages are that the design and business-plan build-out are light, the sales vendor supports the application, and it is accepted on a rolling application basis.
The eligible items center on products that work for routine tasks on physical sites, and the published categories include the following.
- Cleaning robots / food-serving robots / beverage-restocking robots (labor-saving in the back-of-house and in customer service)
- Ticket-vending machines / self-checkout machines / automatic check-in machines (self-service for stores and facilities)
- Other general-purpose labor-saving products in the IoT/robot family
The rough subsidy ceilings as of the application guidelines (the brackets after the March 19, 2026 revision) are as follows. The figures in parentheses are the ceilings when the pay-raise special provision is met.
| Employees | Subsidy ceiling (standard) | Subsidy ceiling (pay-raise special) |
|---|---|---|
| 5 or fewer | 200万円 | 300万円 |
| 6–20 | 500万円 | 750万円 |
| 21 or more | 1,000万円 | 1,500万円 |
The subsidy rate is 1/2 or less. For labor productivity, you are required to "improve it by a compound annual growth rate (CAGR) of 3.0%以上 each year for the three years after the subsidized project ends, compared with the time of application."
There is one fact a buyer must always recognize here. The catalog type's ceilings were lowered in March 2026 (before the revision, even the "5 or fewer" bracket was on the order of 500万円). Trusting the values in "last year's article" at face value will cause accidents. When the amount is the deciding factor in an investment decision, verify the application guidelines at the time of application against the primary source.
The catalog type suits the case where "what you're installing is decided, and it's in the catalog." If you have almost no custom requirements and want to prioritize immediate impact, it is the first candidate.
4. The general type — made-to-order system construction, AI, and automation
The general type is a track that supports equipment introduction and system construction tailored to your own site and business content. This is the real target for a buyer considering labor-saving through AI and business-system development.
That is because the general type's eligible expenses are defined as follows, and system construction costs are "mandatory."
Machinery and system construction costs (mandatory), technology-introduction costs, expert costs, transportation costs, cloud-service usage fees, outsourcing costs, and intellectual-property-related costs
In other words, the program's scope includes the cost of outsourcing new system development, cloud usage fees, and expert costs. Moving operations such as order-taking, inventory, and dispatch onto a system built to your own specifications, or embedding AI to automate judgment, data entry, and reconciliation — this kind of software-centric labor-saving rides exactly on this expense category.
The rough subsidy ceilings (as of the application guidelines) come out an order of magnitude larger than the catalog type.
| Employees | Subsidy ceiling (standard) | Subsidy ceiling (pay-raise special) |
|---|---|---|
| 5 or fewer | 750万円 | 1,000万円 |
| 6–20 | 1,500万円 | 2,000万円 |
| 21–50 | 3,000万円 | 4,000万円 |
| 51–100 | 5,000万円 | 6,500万円 |
| 101 or more | 8,000万円 | 1億円 |
The subsidy rate is 1/2 for SMEs (2/3 if you carry out a large pay raise), and 2/3 for small-scale businesses and businesses undergoing rehabilitation.
On the other hand, the requirements become clearly heavier. As of the application guidelines, targets are imposed such as, for pay raises, "increase the CAGR of total salary paid per person by 3.5%以上" (a combined 6.0%以上 for a large pay raise), and, for labor productivity, "increase by a CAGR of 4.0%以上." Because these levels move each application round, always confirm the figures against the latest version. The essence to grasp here is that the general type is a trade-off: "in exchange for being able to invest big, you are required to commit to productivity and pay raises."
The general type suits the case where you rebuild your own company-specific operations — the ones not listed in the catalog — with a system to save labor. Its strengths are the large ceiling and the fact that software development costs, cloud costs, and outsourcing costs can be squarely eligible.
5. Which type does labor-saving through AI and systems ride on?
This is where buyers hesitate most. The decision is settled almost mechanically along two axes: "is the target of investment in the catalog?" and "is custom development required?"
省人化したい業務がある
│
▼
その解決策は「既製の汎用製品」で足りるか？
├─ YES（券売機・配膳ロボット等、カタログに存在）
│ └─▶ カタログ注文型（軽い手続き・即効・販売事業者がサポート）
│
└─ NO（自社業務に合わせた設計・開発が必要）
│
▼
システム構築・AI自動化・DXが中心か？
└─ YES ─▶ 一般型（機械装置・システム構築費が必須経費に乗る）
To give concrete examples—
- A restaurant introducing a ticket-vending machine or food-serving robot → catalog type. Choose the product and apply with the sales vendor.
- Moving order-taking, billing, and dispatch onto a SaaS built to your own specifications → general type. System construction costs, outsourcing costs, and cloud costs are eligible.
- A business system that automates inquiries and document reconciliation with AI → general type. Being custom development, it doesn't exist in the catalog.
The B2B subscription SaaS for the lumber-distribution industry that I worked on was exactly the type of labor-saving where "order-taking and logistics coordination that had been running on phone, fax, and paper are moved onto a system built to the company's own specifications, reducing person-dependent work." Labor-saving through this kind of made-to-order system construction is, by its nature, the world of the general type, not the catalog type. I've organized the axes for deciding which technology to build with in a technology-selection framework for legacy-industry DX, but think of choosing the subsidy type as the step "one level above" that technology selection — deciding first whether an off-the-shelf product is enough at all, or whether you build it out.
6. How to think about measuring the effect of labor-saving
Both to meet the subsidy requirements and to avoid a bad investment decision, it is decisively important not to talk about the effect of labor-saving in terms of "hours saved." What the program measures is labor productivity, and the buyer should design to match.
Labor productivity is captured roughly in the following form (follow the application guidelines for the precise definition and formula).
労働生産性 = 付加価値額 ÷ 労働投入量（従業員数 × 労働時間）
付加価値額 ≒ 営業利益 + 人件費 + 減価償却費 ※制度上の定義に従う
目標： 補助事業後3年間、この値を年平均成長率(CAGR)で
カタログ型 3.0%以上 / 一般型 4.0%以上 向上させる
From here, the effect-measurement design a buyer should do before applying comes down to the following three steps.
- Take a baseline — capture, in numbers, the "person-hours that operation consumes" and the "value-added it produces" before introduction. Without this, you cannot claim an effect later.
- Break down how the labor-saving works — with the system/robot introduction, "which task's person-hours, and how many" are cut. Is it designed so the cut person-hours shift to higher value-added work (i.e., reduce the denominator without dropping the numerator)?
- Translate into KPIs and CAGR targets — make sure you can explain on a single page how on-site KPIs like person-hours saved, number of cases processed, and lead time ultimately connect to the labor-productivity CAGR.
Failures in practice usually stem from "talking about the effect only as time reduction and being unable to explain the value-added numerator." Even if input automation frees up 40 hours a month, if those 40 hours produce nothing, productivity does not rise. Only when you design where the freed-up time goes do both the program requirements and the investment return hold up.
Note that with labor-saving through systems, not only the "it got faster" effect but also the "mistakes decreased" effect contributes to value-added. On the payment platform I was involved in, by building out replay control and idempotent design, we achieved zero double charges in production. The person-hours and losses that would otherwise have gone to after-the-fact handling of double charges, refunds, and reputational damage vanish entirely — this is a textbook case of "labor-saving through quality," and it works on both the numerator and the denominator of the effect measurement. If you frame labor-saving not just as "making work faster" but design it to include "eliminating rework," your ability to explain return on investment rises a notch.
7. What a buyer should decide before applying
Once you've chosen the type, there are points the buyer's side (i.e., the business receiving the subsidy) should take the lead in nailing down. These are the parts you cannot dump entirely on a vendor or a support organization.
- The target operation for labor-saving and its current values — which operation's person-hours you cut, how many, and why. Only your own company holds the baseline (Section 6).
- The scope of requirements and the basis for the estimate — the general type questions the validity of the estimate and business plan. Get to a state where the buyer can explain "why this feature costs this much."
- The maintenance and operations structure — in principle the subsidy covers up to introduction, but for a system the real game begins after it's installed. If operations stumble, you can't meet the productivity target (the 3-year CAGR).
- Schedule alignment — the catalog type has rolling intake and the general type is round-based, and as a rule, if you place the order or start work before the grant decision, it becomes ineligible. Reconcile the ordering between the development lead time and the application schedule early.
The subsidy is a program that "returns part of the investment later"; it is not a program that takes over the investment decision itself. Whether the ROI of labor-saving is reasonable even without the subsidy — whether the buyer holds that core is what separates success from failure.
8. Common failures (labor-saving anti-patterns)
Finally, here are the typical ones to avoid from the buyer's viewpoint.
- Deciding on last year's numbers — the subsidy rate, ceilings, and requirements are revised every year. In fact, the catalog type's ceiling dropped in March 2026. Always look at the application guidelines at the time of application.
- Choosing the wrong type — trying to push custom development through the catalog type, or building a heavy plan under the general type when an off-the-shelf product would suffice. It's decided mechanically by the nature of the investment target (Section 5).
- Talking about the effect only as time reduction — not designing where the freed-up person-hours go, and being unable to explain the labor-productivity numerator.
- Over-investing because the subsidy exists — trying to use up to the ceiling and piling on features you don't actually need. The self-funded portion after the subsidy (1/2 to 1/3) is real money out of pocket.
A note on neutrality: The author is not an IT-introduction-support vendor and does not file subsidy applications on your behalf. This article is a neutral explanation to support a buyer's decision-making. For the details of the program and the latest subsidy rates, ceilings, schedule, and requirements, always check the official application guidelines (the SMRJ "SME Labor-Saving Investment Subsidy" site). The value this article can provide is not "application support" but the design of the buyer's-side judgment: which labor-saving investment is reasonable for your company, and how to measure its effect. As a first step, I recommend starting with a free consultation / DX assessment to inventory the target operations for labor-saving and their baseline.