In Japanese Care Software, One User Means Three Different Things

What a single July service-sheet batch costs at Kaipoke, CareManager-kun and WizMan, priced per head, per ID and per office

SeaPRwire – It is the third week of July 2026 and a care manager at an office in Saitama is sending service provision sheets to 20 home-care providers. Under the old workflow this took most of an afternoon: print, envelope, stamp, repeat. Under the Care Plan Data Linkage System, which the health ministry pushed hard in the June 2026 fee revision, it takes 15 to 20 minutes for the same 20 offices, using bulk send.

That is the good news, and it is real. The bad news starts when the office asks which software to run it on, because the three vendors that dominate the Japanese care market bill in three different units, and at ten care managers the difference between the two published options is a factor of 2.1.

This is a comparison built around that one task rather than around a feature matrix, and it is priced in yen at 2026 rates because that is the only number a manager can act on.

The task: one batch, twenty offices, one deadline

Set the task precisely, because care software comparisons fail when the task is vague.

Every month a care management office sends a service provision sheet to each provider that served its clients. The sheet says who received what service on which days. Providers use it to bill the long-term care insurance system. The care manager uses the return to reconcile. In July the batch is larger than usual, because the June fee revision and the August meal-cost adjustment both land inside the same fiscal year.

Since the June 2026 temporary revision, the treatment improvement add-on, known as shogu kaizen kasan, carries a special condition: the office must use or formally pledge to use the Care Plan Data Linkage System. That single sentence moved adoption further than a decade of encouragement had. Adoption reached 53.1% of offices as of the end of April 2026, according to health ministry figures presented to the Social Security Council, against roughly 40% reported by operators in July 2026.

The two figures are not in conflict. They are different dates and different denominators: the ministry counts offices registered, operators count offices actually exchanging data. The gap between registration and use is where most of the practical problems live.

What the June revision did to the task

Two changes matter for anyone choosing software in the fourth quarter of 2026.

The first is money. The 2026 temporary revision added 2.03% to care fees, split into 1.95% applied from June 2026 for treatment improvement and 0.09% from August 2026 for the higher standard meal cost. The core measure is a 10,000 yen monthly pay increase for care staff, plus 7,000 yen a month extra for offices that take part in productivity and collaboration programmes.

That 7,000 yen line is why software budgets are being reopened. Productivity is now being paid for, and software is the cheapest way to evidence it. It is also the smallest possible lever: an office with ten care managers is chasing 70,000 yen a month in add-on revenue, which is roughly the cost of two mid-range subscriptions.

The second change is a deadline disguised as a feature. The Care Plan Data Linkage System is scheduled to be absorbed into the national care information platform in January 2027, at which point data exchange becomes a free function of the Care Web Service. Offices that have not registered by then lose the ability to exchange data at all. From January 2027, the linkage itself costs nothing.

Put those two facts together and a procurement rule falls out. In the fourth quarter of 2026, do not buy a care software package because of its data linkage module. You would be paying for something that becomes a free government service within months. Buy for records, for billing, or for the workflow around them.

The bill behind the bill

It helps to know why the revision was only 2.03% and why no office should expect another one soon, because that constraint is what makes software the only margin lever left.

Long-term care insurance spending reached 12.22 trillion yen in fiscal 2025, with 5.78 million recipients, according to health ministry figures reported by Nippon.com in 2026. By April 2026 the monthly certification count stood at 7.61 million people, with 4.84 million receiving benefits and monthly service cost at 1.009 trillion yen.

Against that, the premium paid by a first-category insured person, someone aged 65 or over, averaged 6,225 yen a month in fiscal 2026, a record high. There are about 36 million people in that category, so the premium line alone is worth roughly 2.7 trillion yen a year and is already at the point where further increases are politically expensive.

An office reading those numbers should draw one conclusion. Fee revisions will be small, eligibility will be reviewed, and the 20% co-payment threshold is under active discussion at the Social Security Council as of 2026. None of that is within a care manager’s control. What is within control is the hours spent on paperwork, and that is where the software decision pays or does not.

Three tools, three billing units

Here is where the comparison gets interesting, because the three products do not price the same object.

Kaipoke, from SMS, charges a flat base per office. The care manager edition runs at 5,500 yen a month including tax, with no setup fee and no support fee, and the published entry point for other service lines is 1,000 yen. Its AI add-on, Kyotaku Mana from Life Shift, is priced per registered care manager: 11,000 yen for one, 22,000 yen for two, 33,000 yen for three to ten, all including tax. Ten care managers cost the same as three.

CareManager-kun, from Japan Care Communications, charges per ID. The parent ID is 7,000 yen a month before tax, each sub-ID is 1,000 yen, and the AI prediction option is 3,000 yen per office number. Cost rises linearly with headcount and never caps.

WizMan, from WizMan System, does not publish a price. It quotes per office, bundles the modules, and reports more than 61,200 offices using its products. There is no per-head rate to compare, which is not an accident: a bundled quote is harder to benchmark than a list price.

Table 1: Monthly cost of the same care management office at four headcounts, in yen including tax

Source: Life Shift pricing note on the Kaipoke integration , August 2026, and Japan Care Communications price list , 2026. WizMan does not publish list pricing

Read the per-head row carefully, because it is the only number that decides anything. At one care manager, Kaipoke costs 1.65 times CareManager-kun. At ten it costs 2.1 times. The ranking does not flip, but the size of the gap does, and it moves in the opposite direction from what a per-seat buyer expects.

The reason is the cap. Kyotaku Mana stops charging at ten registered users, so the marginal care manager from four to ten is free. CareManager-kun charges 1,000 yen for every one of them. An office planning to grow from three to eight care managers should model that curve before signing, because the crossover in cost per head sits around the fifth hire.

Now put the difference in annual terms. An office with six care managers pays 38,500 yen a month on Kaipoke with the AI add-on, and 15,000 yen on CareManager-kun with the AI option. The gap is 23,500 yen a month, or 282,000 yen a year.

That number is worth naming in the unit a manager understands. It is about one and a half months of a care manager’s salary, and it is roughly the cost of replacing one aging tablet per staff member every three years. It is not nothing and it is not decisive. Which is exactly why the decision should be made on the workflow rather than the invoice alone.

What each one does at each step

Cost is only half the comparison. The other half is what happens between opening the file and getting confirmation from 20 providers.

Table 2: The July batch, step by step, against what each tool documents

Source: Kaipoke product documentation , CareManager-kun specification notice , and WizMan home care product page , all accessed October 2026

On paper the three are equivalent. That equivalence is the least useful finding in this article and the most commonly published one, which is why feature matrices should be ignored.

The differences show up in the exceptions, and they show up in the order the exceptions occur.

The tool that will not quote

WizMan deserves its own section because its absence from the price table is itself a finding, and because more than 61,200 Japanese offices run its products. That installed base is not an accident of pricing. It is the residue of a distribution model built on regional dealer networks and long relationships, and it shapes what the product does well.

The advantage is network density. WizMan’s home care system shares usage and provision data directly between offices that both run WizMan System SP, inside the same corporation or across contracted partners. In a prefecture where the local hospital group, the visiting nurse station and half the home helpers all run the same product, that is a shorter path than a national standard specification, because there is no translation step between two vendors’ interpretations of the standard.

That is precisely where the back-side transmission failure comes from in mixed networks. Two offices on the same product interpret the specification the same way. Two offices on different products sometimes do not.

The disadvantage is benchmarking. A bundled quote covering records, billing, transmission, tablet access and support cannot be compared to a 38,500 yen subscription without unbundling it, and vendors know that most buyers will not unbundle. The way to do it is to ask for a quote on the narrowest possible scope, care management records plus transmission plus 20-provider sharing, and then compare that number to Table 1. If the narrow quote lands within 30% of the published options, the bundle is probably fair. If it lands at three times, you are buying modules you will not use before the January 2027 migration makes half of them redundant.

One more question is worth asking of any bundled vendor in 2026. What happens to the contract when the linkage function becomes free? A vendor with a long installed base has an interest in migrating customers to the new platform smoothly, and a vendor with a thin one has an interest in locking the renewal before the change. The answer tells you which kind you are talking to.

Where it breaks

Operator accounts from July 2026 describe the same failure in more than one office. The service provision sheet has a front side and a back side. The front side transmits. The figures on the back-side attachment do not, because the transmitting software and the receiving software implemented the standard specification differently, or because the receiving office runs a product that does not support that part of the specification.

The workaround now repeated across the country is a single sentence: you can import what we sent, but please re-check it.

That is a small failure with a large cost. The entire point of the linkage system is to remove the re-keying step. A partial transmission removes the printing and keeps the checking, which captures maybe half the time saving and all of the trust problem. It also means the 15-to-20-minute batch is not the end of the task but the middle of it.

The mitigation available to any office in October 2026 is unglamorous. Before the first send of each quarter, run one test file to each provider and ask them to confirm that the back-side figures arrived. Twenty providers means twenty test files, once. That is an hour that saves a month of reconciliation.

The second failure mode is timing. Anything an office buys in the fourth quarter of 2026 for data linkage has a useful life measured in months, because January 2027 brings the platform migration and the free web service. Vendors know this, which is why the fourth-quarter pitches emphasise records, AI drafting and billing rather than linkage.

The third failure is arithmetic, and it is the one nobody puts in a case study. A bulk send to 20 offices is fast because the destination is inferred from the sheet. When the inference is wrong, the file goes to the wrong provider, and the provider has no way to know it should not have received it. The correction costs a phone call and a resend, which is fine once and corrosive at scale. Offices sending to more than 40 providers should build a manual confirmation list rather than trusting the auto-detection for every name.

Pick by invoice, not by feature list

If you are choosing in the fourth quarter of 2026, the decision tree is shorter than the brochures suggest. Work through it in this order.

• Count the care managers you will have in April 2027, not the ones you have now. Every product here prices differently at five heads than at ten, and the crossover between the two published options sits near the fifth hire.
• Decide whether you are buying drafting help or billing compliance. AI drafting is the line item that grows. Billing and transmission are the line items that are already commoditised. Do not pay an AI premium for a transmission feature.
• Check what your providers actually run. A linkage system is only as good as the receiving end, and the back-side failure is a receiving-end problem you cannot fix by buying better software.
• Ask what happens in January 2027. Any contract running past that date should state, in writing, what the vendor charges once the linkage function becomes free.
• Do the subsidy paperwork for hardware and setup only. The care technology introduction support programme covers up to about three quarters of introduction costs and does not cover monthly fees. Every recurring line in Table 1 is yours.

The last point catches small offices. The subsidy is generous exactly where the cost is smallest, the one-time purchase, and absent where the cost recurs, the subscription. An office that plans around the subsidy will find it funds the tablet and not the software.

There is a sixth consideration that decides more outcomes than any of the above, and it has nothing to do with the vendor. Who in the office will actually own the file format. A tool is only as fast as the person who maintains the template, the provider list and the naming convention. In an office of three care managers that person is usually the manager, which means the software decision is really a decision about whether the manager has two hours a month to spare.

What to do this week

One action, and it takes an afternoon.

Export last month’s provision sheet batch from whatever system you run now, send one file to the three providers you exchange the most data with, and ask each of them one question: did the back-side figures arrive, and did the file import without manual correction.

The answers will tell you more about your switching cost than any demonstration. If all three say yes, your current system is fine and the question is price. If two say the figures were missing, you are paying for a transmission step you are still doing by hand, and that is the line to renegotiate at the January 2027 renewal rather than the feature list.

Methods

Data as of October 2026. This is a specification-level comparison against published pricing and documentation, plus one operator’s published account of running the July 2026 batch across 20 provider offices. It is not a live trial: no resident data was processed, and no vendor provided a demonstration environment. Where a claim depends on hands-on use, the text says so and attributes it to the operator account rather than to this evaluation.

Prices come from vendor price lists and press notices published in August and September 2026, converted to tax-inclusive where the source quotes tax-exclusive. Kaipoke’s 5,500 yen base is the care manager edition; other service lines start at 1,000 yen and are not comparable. CareManager-kun’s 7,000 yen is a parent ID before tax, so the table adds 10% consumption tax and the 1,000 yen per sub-ID. The AI option is 3,000 yen per office number rather than per user, which is why it appears as a flat addition.

Two caveats. WizMan’s absence from the cost table is not a judgement about value, it is the absence of a published rate, and any comparison that fills that cell with an estimate is inventing a number. And the adoption figures of 53.1% and 40% come from different sources counting different things, so neither should be used to argue that adoption is strong or weak without naming which one is meant.

Sources

• Japan Care Communications, CareManager-kun pricing and FAQ , per-ID pricing and AI option, 2026
• Life Shift pricing note via SEO Academy , Kaipoke and Kyotaku Mana combined monthly cost, August 2026
• WizMan, home care management product page , module coverage and MeLL+ integration, 2026
• Social Security Research Institute, 136th Long-Term Care Insurance Subcommittee , Care Plan Data Linkage System adoption at 53.1%, April 2026
• CB Healthcare, integration into the care information platform , January 2027 migration timetable, August 2026

This article was prepared by the expert team at Global Infinity Labs, drawing on insights and data provided through its MCP data services.

About Global Infinity Labs

Global Infinity Labs is an AI innovation company specializing in MCP data services.

Email: support@globalinfinitylabs.com