Money Forward,Inc. Q4 FY2025 earnings call

• Company Overview & Core Business Model
  • Hmcomm is a spin-out venture from Japan's National Institute of Advanced Industrial Science and Technology (AIST), focused on solving social issues through proprietary sound × AI technology. The company name comes from its core mission of 'Human Machine Communication'.
  • The business operates on a two-pillar model: custom co-creation AI solutions that convert into subscription-based AI products over time. Approximately 85% of sales are direct to customer, 15% through partner sales agencies.
  • The company follows a unique R&D-driven business cycle: early R&D → co-creation PoC and joint development with corporate partners → product commercialization → customer adoption and ongoing product improvement, repeating this cycle to drive sustained growth. End-to-end capability from early R&D to social implementation is a core competitive strength.
• Key Operational Highlights for 2025
  • Completed two business acquisitions during the fiscal year: acquired the IT consulting business of IP Partners Co., Ltd. to add upstream consulting capabilities for LLM implementation that were previously lacking, and acquired a portion of business from Fantaractive Co., Ltd. to complement UI/UX and design capabilities.
  • Signed an agency agreement with Sojitz Tech Innovation Co., Ltd. to expand sales of the company's flagship generative AI product Terry2, as part of the broader strategy to grow sales through partner agency channels.
  • Launched FAST-D, an abnormal sound detection product that combines the company's core sound technology with satellite data for infrastructure DX projects, and started a joint project with a local Japanese government.
  • Advanced social implementation of Terry2, a generative AI conversational agent that combines real-time API and LLM technology, with financial institution clients.
  • The number of AI Solution projects and AI Product accounts both grew steadily, and average prices showed an improving trend.
• Financial Position
  • All interest-bearing debt has been repaid, equity ratio exceeds 80%, creating a very solid financial base.
  • Cash and cash equivalents remain at approximately 60% of total assets even after two business acquisitions, leaving ample investment capital for future growth.
  • Operating cash flow is positive; investment cash flow is negative due to acquisition spending, and financing cash flow is negative due to treasury stock purchases.
• Growth Strategy
  • To meet the new Tokyo Exchange Growth Market listing requirement of 100 billion yen market capitalization within 5 years, the company is pursuing two core growth strategies: expand organic growth of existing business (increasing the revenue share of AI Products), and drive non-continuous growth through M&A, with all M&A activity required to contribute to organic growth.
  • For organic growth: accelerate the conversion of completed co-creation AI Solution projects into AI Products, expand sales agency partnerships for AI Products, and drive cross-selling of solutions and products to new customers acquired through agency channels, with growing AI Product revenue as the core intermediate goal.
  • M&A and capital alliance policy based on SWOT analysis:
    1. Expand into adjacent areas of sound AI expertise, and strengthen overall AI capabilities through capital participation or M&A in non-sound AI areas.
    2. Pursue business turnaround of IT development companies that can double or triple their operating scope through partnership with Hmcomm.
    3. Acquire talent and capabilities to address the ongoing industry-wide talent shortage.
    4. Strengthen the recruitment foundation to create access to high-level talent that was difficult to reach with traditional approaches.
  • Established the M&A/PMI Promotion Office with dedicated experts under the direct control of the CEO to systematically source targets and execute on M&A strategy.

For the 2025 December fiscal year, total company revenue was 1.11 billion yen, with operating profit of 0.03 billion yen and an operating profit margin of 3.6%. The company operates two core segments: 1) AI Solutions: This segment contributes approximately 60% of total revenue, came in at ~0.666 billion yen for the period, performed in line with plan. Both the number of projects and average project price have grown steadily year over year, with an average project price of 4.5 million yen. 2) AI Products: This segment contributes approximately 40% of total revenue, came in at ~0.444 billion yen for the period, performed slightly below plan. The number of product accounts and average revenue per account are both growing, with an average annual revenue per account of approximately 10.2 million yen. There is significant variation in project/account size, with many projects/accounts exceeding 10 million yen in value. Customer concentration risk has reduced steadily over time, with more diversified customer base lowering overall business risk.

• For the 2026 December fiscal year, management expects more than 20% year-over-year growth across both AI Solutions and AI Products, driven by the conversion of resources from the 2025 acquisitions into revenue. The full-year guidance is:
  • Total revenue of 1.37 billion yen, representing 23.5% year-over-year growth
  • Operating profit of 0.04 billion yen, with an operating profit margin of 3.3%
  • EBITDA is expected to reach 0.1 billion yen, showing an increasing profit trend even after accounting for goodwill amortization from acquisitions
• Management has committed to achieving a 100 billion yen market capitalization within 5 years to meet the updated Tokyo Exchange Growth Market listing maintenance standards, and has built out dedicated organizational structure to deliver on this target.

• The company faced significant difficulty securing sufficient skilled talent in 2025, which required reliance on external partner resources and contributed to higher outsourcing costs and lower-than-planned profit for the fiscal year.
  • AI product revenue came in below plan for 2025, as the conversion of AI solution projects to products takes longer than initially expected, leading to a temporary higher revenue share for AI Solutions.
  • Goodwill amortization from the 2025 acquisitions created a short-term drag on profit, leading to full-year profit missing the original plan.
  • High precision (near 100% accuracy) is required for public infrastructure and security use cases, which creates high technical barriers to commercialization that can delay project timelines.

• Q: What recent trends and growing demand are you seeing in the AI Solutions business? How do you acquire new customers, and have inquiries been growing?

A: Inquiries have increased sharply recently as many corporate clients are confused about how to integrate generative AI and LLM into their existing operations, and most inquiries are from clients who want to adopt AI but do not know how to implement it. Before listing, customer connections came primarily through AIST collaboration during fundraising. Now, with more successful social implementation cases, PR through timely disclosure and press releases leads to many inbound inquiries from new potential partners, and these inquiries have been growing steadily.

• Q: How much variation is there in revenue per project/account? Do you have projects over 10 million yen? Is there enough talent to handle growing AI Solutions demand, and what is your approach to hiring and training?

A: The average project price for AI Solutions is 4.5 million yen, and the average revenue per account for AI Products is ~10.2 million yen, but there is very large variation by client and product, and yes, there are many projects and accounts exceeding 10 million yen. 2025 was a very challenging year for talent recruitment. We have addressed shortages by entering a business alliance with a Vietnamese firm to utilize their development resources. We also established a cross-functional organization to handle the growing volume of AI-driven development including generative AI and LLM, and we run company-wide OJT programs to allow AI-inexperienced new hires to build expertise, knowledge and practical experience through project participation.

• Q: Can you share quantitative impact examples of AI product implementation, and how does your Japanese-specialized speech recognition accuracy compare to competitors?

A: For example, our Voice Contact product for call centers performs real-time speech-to-text transcription of customer calls, then automatically summarizes the call log for CRM systems using generative AI. This reduces post-call work time, cuts average call handling time to 1/2 to 1/3 of the original level, and allows clients to reduce the number of required operators by 1/2 to 1/3. We have built specialized Japanese speech recognition expertise starting from AIST technology over many years, and we support specialized terminology and custom tuning that general global cloud speech recognition services cannot provide. We also own our proprietary speech recognition engine and support on-premise deployment, which many competitors do not offer, and benchmark testing generally shows our performance is superior to competing offerings in Japanese specialized use cases. The main unique difficulties for Japanese include homonyms and dialects, and the company is equipped to handle these challenges better than generalist providers.

• Q: What future does Hmcomm envision for its Human Machine Communication mission, and what social issues and fields do you aim to support?

A: We see growing social need for solutions to aging social infrastructure, highlighted by recent high-profile water main break incidents in Japan. We aim to work with partners that have complementary non-sound technology to solve these social infrastructure issues proactively, and build more comfortable, safe communities. We also focus on areas like disaster prevention, crime prevention, and monitoring, where shrinking working populations make it increasingly difficult to allocate sufficient human labor. For example, we previously ran a joint project with a major security company for NEDO that used sound AI to automatically detect fights or car accidents from surveillance camera audio to trigger automatic alerts. This field requires near 100% accuracy to avoid false alerts that waste security personnel time, and while the original project is not currently active, we plan to revisit this space as our technology advances and social need grows, because our technology can provide meaningful solutions to these labor shortage problems and help protect Japan as a society.

• Q: [Unanswered/unpublished additional Q&A topics listed in the transcript: expected growth areas, Pasona Masters course initiatives, role of Terry2 during peak inquiry surges, M&A and capital alliance strategy, future IR activities] No full transcripts of these exchanges are included in the provided document.