Toolso.AI
Toolso.AI
All ToolsCategoriesTrendingLatest ToolsPricingBlog
Toolso.AI
Toolso.AI
Toolso.AI
Toolso.AI

Discover the best AI tools to boost your productivity

GitHubGitHubTwitterX (Twitter)YouTubeYouTubeTikTokEmail

Popular Categories

  • AI Writing
  • AI Image
  • AI Video
  • AI Coding
  • More Categories

Explore

  • Latest Tools
  • Popular Tools
  • More Tools
  • Submit Tool
  • Pricing

About

  • About Us
  • Contact
  • Blog
  • Changelog

Legal

  • Cookie Policy
  • Privacy Policy
  • Terms of Service
  • Refund Policy
© 2026 Toolso.AI All Rights Reserved
Limited timeLimited-time offerFeatured Listing24h priority review · No backlink · 30 days featured$29.90then $59.90Price rises to $59.90 after Oct 31Ends in--:--:--Submit now
  1. Home
  2. All Tools
  3. Business Tools
  4. HEROZ
HEROZ interface preview
HEROZ logo

HEROZ

HEROZ is a Tokyo-listed Japanese AI company whose core technology came out of championship shogi engines. Its main commercial product today is HEROZ ASK, a corporate generative AI platform that puts Claude, Gemini and ChatGPT behind one governed company entry point.

Business ToolsAI AssistantGenerative AI Platform#Enterprise#Translation#Machine Learning
Try for Free
Saves
Visits
Views
Pricing
Freemium
Published
Aug 22, 2026
Domain
heroz.jp
Community rating

Used this tool? Rate it

Rate this tool

HEROZ Product Information

Try for Free
Tool Information
Saves
Visits
Views
Pricing
Freemium
Published
Aug 22, 2026
Domain
heroz.jp
Community rating

Used this tool? Rate it

Rate this tool

Featured Tools

Related Tools

Try for Free

What is HEROZ?

HEROZ is not a consumer AI tool you sign up for. It is a listed Japanese AI company, and the entry that brought you here describes the company rather than a single product. Understanding that distinction is the first useful thing to know about it: the site has no registration form, no free tier, and no product dashboard — its top-level navigation runs to services, company information, news, investor relations, recruitment and contact.

The company describes its own core technology as machine learning methods, deep learning included, accumulated through the development of shogi AI and now applied to problems across many industries rather than to board games alone. On its English pages the company gives that core technology a name — HEROZ Kishin — and traces it to AI development for Japanese chess, chess, backgammon and go. That origin is not marketing colour. The company's engineers built engines that beat professional players at a game whose branching factor dwarfs chess, and the machine learning practice developed in that pursuit became the asset it now sells to banks, construction firms and manufacturers.

ITmedia reported that the Tokyo Stock Exchange approved the listing on 15 March 2018, with trading beginning on 20 April 2018 under securities code 4382, and that the company was founded in 2009. Being publicly listed changes what can be known about it: unlike most entries in an AI tool directory, its revenue, margins and strategy are disclosed on a legal schedule rather than asserted in marketing copy.

For a reader who actually wants to use something, the practical answer is HEROZ ASK — the company's corporate generative AI platform, and the subject of most of the product detail below.

One correction worth noting: the address recorded in many directories, heroz.jp, redirects to a different domain, heroz.co.jp. Only the latter carries content, and everything here was verified against it.

Core Features

  • HEROZ ASK as a governed multi-model front end: The product page states that it supports the latest models from Claude, Gemini and ChatGPT, which places it in the category of a model-agnostic corporate front end rather than a proprietary foundation model. The value proposition is not a better model; it is one controlled doorway to several, with company policy and company data attached.
  • Enterprise data retrieval, summarisation and translation: The platform's stated purpose is to search, summarise and translate the material an organisation already holds. This is the workload that distinguishes a corporate deployment from individual chatbot use — the answers are supposed to be grounded in internal documents, not in the open web.
  • A functional bundle aimed at corporate requirements: Official materials describe generative chat, internal data utilisation, security, slide generation and external system integration as a single package, on the reasoning that companies stall when an AI tool sits isolated from their existing systems and staff resort to copying and pasting between them.
  • Voice and meeting material processing: Transcription of audio into usable meeting records is presented as one of the routine tasks the assistant absorbs, alongside document handling.
  • In-house machine learning research, not only model resale: The company states that its engineering bench includes the developer of a shogi AI that defeated an active professional player and a developer from the research arm of a major IT company. It also maintains a dedicated artificial intelligence research department, and holds membership in Japan's principal deep learning and artificial intelligence professional bodies.
  • Industrial AI delivery beyond the SaaS product: Alongside the packaged platform, the company builds applied AI for specific industries. Its own materials note that its AI is used well beyond board games, including at major financial institutions.

Use Cases

  1. Consolidating scattered AI usage inside a company: The most common corporate problem this class of product addresses is not the absence of AI but its uncontrolled presence — staff using personal accounts on assorted services with company documents. A single governed platform with organisational billing and security settings replaces that sprawl, which is why the product leads with security as standard rather than as an upgrade.
  2. Making internal documents answerable: Connecting the assistant to the organisation's own storage turns a general chatbot into something that can answer questions about your contracts, specifications and reports. This is the difference between a novelty and a tool people return to.
  3. Routine document work at departmental scale: Summarising long material, translating between Japanese and other languages, drafting slides and turning recorded meetings into written records are the tasks that consume the most junior-analyst hours in Japanese corporates, and they are the ones the platform is positioned to absorb.
  4. Industry-specific AI projects: For organisations whose need is not a chat interface but a model applied to their own operational problem — demand forecasting, inspection, risk scoring — the company's business is delivering that as a project. Its consulting-style B2B work, rather than the SaaS product, is what its financial institution and construction clients bought.
  5. Consumer brain-game AI: Its consumer side runs on smartphones: a shogi app officially recognised by the Japan Shogi Association, and a mobile board game co-developed with The Pokémon Company. These remain live products and part of the company's identity, though they are no longer where the growth story sits.

How to use HEROZ

  1. Decide which HEROZ you actually need. The company sells a packaged platform and it sells bespoke industrial AI. These are different purchases with different timelines, and the website separates them accordingly.
  2. For the platform, start at the product site. HEROZ ASK has its own domain and its own materials, distinct from the corporate site. That is where pricing structure, feature lists and integration details live.
  3. Book a consultation rather than expecting self-service. There is no public self-signup flow of the kind a consumer tool offers. The stated route is a free consultation about AI adoption, which is normal for Japanese enterprise software procurement.
  4. Scope a department, not the whole company. Because the pricing model shares an allowance across a team rather than charging per person, the natural unit of a first deployment is one department with a real, repeated document workload.
  5. Connect a data source early. The platform's advantage over a generic chatbot only materialises once it can read your own material. A pilot that never connects internal storage tests the wrong thing.
  6. Verify the security and data-handling terms in writing. Marketing pages describe advanced security as standard; the specifics that matter to a compliance review — retention, model-side data handling, regional processing — are contract questions, not website questions.
  7. For investors rather than buyers, read the IR materials. As a listed company, its filings carry information no vendor page would volunteer.

Tips & Best Practices

  • Treat the shogi heritage as evidence of engineering depth, not as a product claim. Beating professional players demonstrated real capability, but the engine that did it is not what you are buying. What carries over is the team and the method, not the software.
  • Pin down what "supports the latest models" means for your contract. Multi-model access is genuinely useful, but model availability and version currency change quickly. Confirm which models are live in your tenancy and what happens when a provider deprecates one.
  • Measure the shared-allowance model against your actual usage curve. Pooled consumption favours organisations with a few heavy users and many occasional ones. If everyone in the department is a heavy user, per-seat pricing elsewhere may work out comparably.
  • Test retrieval quality on your worst documents, not your best. Scanned PDFs, tables and inconsistent internal terminology are where enterprise retrieval fails. A pilot on clean, well-structured files will mislead you.
  • Ask for reference customers in your own industry. The company's public credibility rests heavily on named enterprise adoption; the useful version of that for you is a reference in your sector, not a logo wall.
  • Read the financial disclosures if this is a strategic dependency. For a listed vendor you can check whether the product line you are betting on is growing or being wound down — information that is simply unavailable for a private AI startup.
  • Do not expect English-language support parity. The company's English presence is investor-facing. Assume the working language of the product and its support is Japanese unless told otherwise in writing.

Who is HEROZ for?

  • Japanese enterprises adopting generative AI under governance constraints: The core buyer is an organisation that needs AI available to staff but cannot accept uncontrolled use of consumer services.
  • Departments with heavy document workloads: Legal, back office, engineering documentation and any team where summarisation and translation consume real hours.
  • Companies wanting applied industrial AI: Organisations whose problem is a forecasting or inspection model rather than a chat interface, and who want a vendor with genuine machine learning depth rather than a prompt-engineering shop.
  • Investors and analysts covering Japanese AI: The listing makes it one of the few pure-play Japanese AI companies with public financials, which is a category of interest in itself.
  • Shogi and brain-game players: A distinct and much larger consumer audience reaches the company through its game apps rather than its enterprise products.

Who it is not for: individuals looking for a personal AI assistant to sign up for; teams needing an English-first product with English-first support; and buyers who require self-service procurement without a sales conversation.

Platforms

  • Corporate website (Japanese): The primary and most complete source, covering services, company information, news and investor relations.
  • Corporate website (English): A reduced set of pages covering service, company overview and IR only — sufficient for an investor or a partner, not for a product evaluation.
  • HEROZ ASK product site: A separate domain carrying the platform's own marketing, feature and pricing materials. This is the practical destination for a prospective buyer.
  • Web-based platform delivery: The corporate product is delivered as SaaS, with integrations into cloud storage and business systems rather than as installed software. (Based on public information)
  • Consumer mobile apps: Smartphone titles including the shogi app and the co-developed Pokémon board game, distributed through ordinary app stores.

There is no publicly documented open API or developer programme for the enterprise platform, and no self-service trial link on the corporate site. Integration is described at the level of connected business systems rather than as a developer surface. (To be verified)

Pricing & Plans

Pricing behaves differently here than for most entries in an AI directory, because two very different things are being sold.

For the packaged platform, HEROZ ASK advertises pricing from 900 yen per month built around a shared allowance for a department rather than per-seat billing, on the argument that per-ID licences waste budget on employees who never log in. That is an unusually specific and unusually low published entry point for Japanese enterprise software, and it is the product's clearest commercial differentiator. One caveat belongs with it: the company's claim to be among the lowest-priced in its category carries a footnote attributing it to the company's own survey conducted in November 2024 — a vendor self-assessment rather than independent benchmarking.

For bespoke industrial AI work, no pricing is published at all, which is standard for project-based enterprise AI delivery.

The company's financial disclosures give an unusually concrete picture of how this converts into revenue. For the fiscal year ended April 2026 the group reported consolidated revenue of 6.424 billion yen, up 8.3% year on year, with operating profit of 523 million yen, up 70.7%. EBITDA came to 1.025 billion yen, up 29.2%. Management disclosed that HEROZ ASK grew 27.8% quarter on quarter in the fourth quarter and passed 200 million yen in annual recurring revenue by the end of April.

Read carefully, those numbers say something useful to a prospective buyer: the flagship platform is growing quickly but is still a small fraction of a group whose revenue comes substantially from elsewhere. That is neither good nor bad on its own, but it is the kind of context a private vendor would never disclose.

Alternatives

  • Microsoft 365 Copilot: The default competitor for any Japanese enterprise already standardised on Microsoft, with far deeper native integration into the documents and mail people already use, and correspondingly less flexibility on model choice.
  • Google Workspace with Gemini: The equivalent option for Google-standardised organisations, with the same trade-off between native integration and vendor independence.
  • Domestic Japanese enterprise AI vendors: Several Japanese vendors sell comparable governed generative AI platforms with local support and domestic data handling, which is often the deciding factor over foreign alternatives regardless of feature comparison.
  • Direct provider enterprise plans: Contracting with a model provider directly gives the newest capabilities soonest, at the cost of building the governance, retrieval and integration layer yourself.
  • Systems integrators for bespoke AI: For the project-based side of the business, the realistic alternatives are Japan's large systems integrators and consultancies, which bring scale and existing relationships but generally less specialised machine learning research depth.

Limitations & Considerations

  • There is effectively no independent review sample. On the Japanese review platform ITreview the product carries a satisfaction score of 4.5 drawn from a single posted review, a sample far too small to support any conclusion about quality. Its category ranking derived from that same single review should be read the same way. No meaningful body of independent user feedback about this platform exists publicly.
  • The valuation-versus-substance question has been asked before, publicly. Toyo Keizai questioned whether the substance matched the valuation, noting a headcount of just 39 employees against a market capitalisation of roughly 140 billion yen. That was written at listing in 2018 and the company has grown substantially since, but it established a scepticism about the gap between AI narrative and business scale that remains a fair lens.
  • The starting business was small. In the fiscal year before listing the company reported revenue of 877 million yen and net profit of 94 million yen, with consumer income coming mainly from in-app purchases. Placed beside the shogi-champion story, that is a useful corrective on how much of the early valuation was expectation rather than trading history.
  • Investor materials are less accessible than they appear. The financial highlights and IR library pages render as empty shells to a visitor, so the numbers require downloading filings rather than reading a page. The figures cited above came from a transcript of the results briefing, not from the company's own summary pages.
  • The English-language surface is thin. English pages cover service, company and IR only. Any non-Japanese organisation evaluating this vendor should assume the working language is Japanese and confirm support arrangements explicitly.
  • The lowest-price claim is self-assessed. The pricing comparison rests on the company's own November 2024 survey, not on third-party benchmarking, and pricing across this category has moved considerably since.
  • Security specifics are asserted, not published. Marketing describes enterprise-grade security as standard, but the public pages do not set out certifications, retention periods, or how data is handled by the underlying third-party model providers. For a platform whose selling point is routing company data to Claude, Gemini and ChatGPT, those are exactly the questions a security review must answer contractually.
  • The company is in the middle of a strategic reshaping. For the fiscal year ending April 2027 the company named two strategic axes: evolving its Variosecure subsidiary into an AI security agent, and creating a new AI BPaaS category around HEROZ ASK Cowork. Active repositioning creates opportunity, but buyers should confirm that the product line they are adopting is one of the ones being invested in.
  • The consumer and enterprise halves can mislead each other. The shogi and Pokémon apps generate the brand recognition; the enterprise platform generates the growth. Evidence about one is not evidence about the other.

FAQ

Q1. Is HEROZ a tool I can sign up for?

Not directly. HEROZ is a company, and its corporate site has no registration or login for individuals. The product a business would actually adopt is HEROZ ASK, which has its own site and a consultation-based sales process rather than self-service signup.

Q2. What is HEROZ ASK exactly?

A corporate generative AI platform. It provides staff with a governed interface to leading commercial models, connects to internal company data for search and summarisation, and bundles the security, slide generation and system integration features that corporate deployments require.

Q3. Which AI models does it use?

Its product materials state support for the latest models from Claude, Gemini and ChatGPT. It is a platform layer over third-party models rather than a proprietary foundation model, and the model line-up in any given tenancy is worth confirming directly.

Q4. How much does it cost?

Published pricing for the platform starts at 900 yen per month, structured as an allowance shared across a department rather than charged per user. Bespoke industrial AI projects are priced individually and no figures are published. The company's claim to be among the cheapest in its class comes from its own survey rather than independent comparison.

Q5. Is the company financially stable?

It is listed on the Tokyo Stock Exchange, and for the fiscal year ended April 2026 it reported group revenue of 6.424 billion yen with operating profit of 523 million yen, both up year on year. Listing means these figures are audited and disclosed on a schedule, which is a materially stronger basis for judgement than most AI vendors offer.

Q6. What does shogi have to do with enterprise software?

The company's machine learning practice was developed building engines for shogi and other brain games, culminating in an AI that beat an active professional player. The engines themselves are not the product; the accumulated method and the engineering team are what the company applies to industrial problems.

Q7. Is the product available in English?

The corporate site has an English section, but it covers company and investor information rather than product documentation. Treat the platform as a Japanese-market product and confirm any English support requirement explicitly before committing.

Q8. What do actual users say about it?

Very little is publicly available. The main Japanese B2B review platform lists a high satisfaction score based on one review, which is not a usable sample. No substantial independent user feedback or technology press evaluation of the platform exists, so buyers should rely on their own pilot rather than on published sentiment.

Q9. Does it still make games?

Yes. The shogi app recognised by the Japan Shogi Association remains a live consumer product, as does the board game co-developed with The Pokémon Company. These are the company's historical foundation and a continuing part of its brand, though the enterprise business is where its strategy now concentrates.

Q10. What should a buyer check before committing?

Three things the website cannot answer: the contractual terms covering data retention and how the underlying model providers handle your content; which models are actually available in your tenancy and what happens on deprecation; and whether the product line is one of the two strategic axes the company has said it will prioritise. All three are procurement questions, and all three matter more than any feature list.


Verified against the company's public pages and its FY2026 results briefing in August 2026. Figures for a listed company change on a disclosure schedule; check the latest filings before relying on any number here.

Know a Similar Tool?
If you know other great AI tools, feel free to submit them to us