[Full Transcript] Money Forward Announces AI Strategy "Money Forward AI Vision 2025"

On April 2, 2025, Money Forward announced its AI strategy, “Money Forward AI Vision 2025.”

“Going beyond operational efficiency with AI”

Money Forward will accelerate the utilization of evolving AI to transition from a DX company to an AX company, with the goal of becoming Japan’s No.1 back-office AI company.

This article provides a full transcript of the press conference.

Tsuji: Hello everyone. I’m Tsuji from Money Forward. Today, I would like to present our AI strategy.

Before that, I must say I recently feel fortunate to be living in this era of AI.  As AI evolves by the day, we’re feeling a mix: half the time it’s a joy to create products, and the other half it’s a struggle to keep up with the rapid pace.

As a startup, we believe we can take on many interesting initiatives.

Introduction

Tsuji: Since our founding in 2012, we have been offering over 60 services, including Money Forward Cloud for corporations, Money Forward ME for individuals, and products for financial institutions.

The number of paying business and sole proprietor customers for Money Forward Cloud has grown from 150,000 in 2021 to over 400,000 recently. We are grateful to see the adoption of our services among businesses and sole proprietors growing at such a rapid pace.

The number of users of Money Forward ME for individuals has also grown from 4.73 million in 2017 to 16.64 million today. We recently announced a joint venture with Sumitomo Mitsui Card Company, Limited.

Accounting firms are also very important business partners for us. Our services are now available to accounting firms across the country.

According to our survey in 2023*, 78% of the top 100 accounting firms by employee size have adopted our cloud-based accounting solution. (*Based on the 2023 edition of the monthly Pro Partner "Top 500 Professional Service Firms (Japanese edition)," as of November 30, 2023, the percentage of accounting firms that have implemented our cloud-based accounting service was calculated.) 

When we first launched our service, many people said, "I'm too scared to entrust my data to the cloud," but I feel that the world has really changed.

We are driven by the vision of "Money forward. Move your life forward."   And this is our official character “me:mo,” who hasn't appeared much before, but you will see frequently in this presentation. 

Now, let me talk about what Money Forward has achieved through the cloud.

With automatic data acquisition via the cloud, we have significantly reduced manual data entry and enabled communication anytime, anywhere. Since our services are not installed systems, automatic updates are performed on our servers, so users do not need to worry about compliance with laws and regulations.

We have been moving from "analog to cloud," and now we aim to shift from "cloud to AI."

As you know, there is a labor shortage, and the back offices of small and medium-sized companies (SMEs) are in a very difficult situation because they are unable to hire people such as accounting staff. We believe that with digital workers, even with fewer staff, high-quality back office operations can be realized.

We aim to empower SMEs and mid-sized companies facing labor shortages with AI and provide digital workers. 

It’s an ambitious goal, but Money Forward aspires to become the No. 1 back-office AI company. We plan to allocate almost all our resources to achieve this goal.

As you know, the AI market is currently experiencing rapid growth. Additionally, we are fortunate to be a leading SaaS company, and our SaaS business continues to grow.

This graph shows the trend of our Business domain's corporate ARR (annual recurring revenue).  We believe we can achieve further growth with AI alongside the growth of our SaaS business.

According to a survey, the global market for AI-powered SaaS is expected to grow at a CAGR of 40.2% from 2024 to 2031, and we believe that we can achieve even greater growth if we become a leading company in the AI market.

By leveraging AI in our SaaS business, we aim to achieve further business growth. We envision a future where Money Forward, which currently provides digital tools, will offer digital workers.

In the future, it will make more sense to invest in digital workers rather than in labor costs.

The current digital tools market size is approximately 2.2 trillion yen, while the digital worker market size is about 13.3 trillion yen, so we are excited about accessing such a large market.

Money Forward's strengths include a large user base of over 16.6 million individual users and more than 400,000 corporate customers, as well as a wide range of over 60 services.   Additionally, we have accumulated extensive business data across our services over 13 years.

We also love creating products, so we have a highly practical product improvement cycle.  Global technology trends are introduced immediately as well.

In terms of development capabilities, the implementation of AI will become increasingly critical. We currently have engineers from over 40 countries, with development centers not only in Japan but also in India and Vietnam, allowing us to significantly increase development speed.

Our goal for AI impact is to more than double productivity and efficiency in all areas, including operations, sales, development, and new features and products utilizing AI. And we are moving forward with the management goal of more than doubling sales per employee in FY2028.

Enhancing Productivity through AI initiatives

Tsuji: Next, let me explain specifically what we are currently doing and planning for the future.

Significantly Improve Productivity of Internal Operations

Tsuji: Several years ago, we created a team called the AI Forward Office, which reports directly to the CEO, to promote AI across the entire company.

There are various initiatives in each job role, but I will introduce three representative examples.

The first is in customer support. As has been mentioned in various places, customer support is an area where AI is a great fit and can be leveraged as a tool.  By building a search database and automating the text generation, we have been able to handle an increasing number of users with the same number of staff.  We estimate that this will reduce the annual expected workload by about 40%, or about 120,000 hours.

The second is in sales. Many of you may already be using it, but AI can automatically take minutes of press conferences like this one. Our sales team also uses AI, with customer consent, to take all minutes of inside sales and field sales meetings and automatically reflect them in customer management tools like Salesforce.

In the past, sales representatives would record notes after meetings, create homework, and make to-do lists, but now that time has been reduced to almost zero. This has led to a reduction of approximately 20,000 hours per year.

The last is in engineering. By introducing environments such as GitHub and Cursor, coding has become dramatically easier. This has increased effective code production per employee by about 36% and will improve further in the future. Here we are describing a reduction of about 310,000 hours of work time, but from another perspective, it means we can create products faster at the same cost. We believe this is an area with a significant impact.

We are also promoting AI initiatives in other job categories.

Security Governance Measures

Tsuji: As we promote AI utilization, security and governance will naturally become extremely important. It is still a big challenge and there are many issues to be addressed.

First, we started by developing internal guidelines for using generative AI. For example, when using external services equipped with generative AI, we are taking precautions from an information management perspective, such as obtaining prior approval from the responsible department, not inputting information entrusted to us through our services. We have also created an internal ChatGPT to ensure that internal information is handled securely and not leaked.

Additionally, we will establish a Data Analytics Governance Committee to discuss governance issues such as access control and security when handling data. With AI evolving at a rapid pace, we need to constantly change and respond accordingly.

Utilizing AI in Existing Products

Tsuji: Next, we will talk about utilizing AI in existing products. Our company CPO, Hirohara, will explain.

Hirohara：Hello. I’m Hirohara, the CPO of Money Forward Business Company. 

From here, I will introduce three features of Money Forward Cloud that utilize AI, along with demo videos.

Money Forward Cloud Accounting Plus for GPT

The first feature is "Money Forward Cloud Accounting Plus for GPT."

This is Money Forward Cloud Accounting Plus, a cloud-based accounting software used by relatively large companies such as listed companies and companies preparing for listing.

In companies of this scale, there is a large volume of data, and transactions are diverse. Yet, they must finalize figures for the closing of accounts within a limited time, which is a complex and challenging task. This feature serves as an assistant to help with such tasks using AI.

Let me show you the demo video.  Similar to using ChatGPT, it starts by entering what you want the AI to do.

In this example, I asked the AI to compare the PL of the previous month and the month before and tell me what the major differences in numbers were. It then highlighted items with significant increases or decreases, like this. If you want to see a further breakdown of SG&A, you can make an additional request, and the AI will provide a deeper analysis of the differences by account title.

What makes this system so unique is that it’s not just about providing numbers; it also predicts and suggests what to investigate next and what actions to take.

During the financial closing period, there is inevitably a lot of work to be done, so this feature has been well received by our clients, who say that it has made the closing process much easier as it checks these matters on their behalf.

Money Forward Cloud Contract: AI Auto-Fill Function

Next, we introduce the AI auto-fill function of our product, Money Forward Cloud Contract.

Money Forward Cloud Contract is a product that provides an electronic contract service, but not only that, it also allows the user to manage a large number of past paper contracts in the cloud. 

To manage paper contracts in the cloud, the user needs to upload paper or PDF contracts to the cloud, digitize them into a format that can be utilized for business operations, and store them. The "AI Auto-Fill Function" handles this work for the user.

For simple data conversion, scanning would suffice, but contracts also contain items customized for each company.  This function enables AI-OCR to extract this data, organize it into a usable format, pick up information for each item, and convert it into data.

Let me show you the demo video for this as well. 

First, upload the contract.  For example, when you upload a lease contract file, the AI analyzes the document. It not only converts it into data but also classifies items and creates a ledger for each item within the system.

For instance, a lease contract includes items like "guarantor," which are not typical in general contracts. One of the key features is its ability to convert such unique, contract-specific items into data.  So in the case of a lease contract, even if you later want to see the "guarantor data," you can do so as it converts everything into data for you.

In addition to lease agreements, there are various other types of contracts, such as business outsourcing agreements and non-disclosure agreements, and the ability to convert all items according to their respective categories into data is a major feature. 

In the past, you had to manually input each contract to convert it into data, but now, with the AI Auto-Fill Function, you only need to convert and check, reducing ledger creation time by about 53%. It is a highly convenient feature.

Money Forward Cloud HRIS: Document Template Auto-Generation Using AI

Lastly, we introduce the AI-powered document template auto-generation feature of Money Forward Cloud HRIS.

Money Forward Cloud HRIS is a cloud-based system that digitizes procedures such as onboarding and offboarding. Here, I will show you a function that automatically generates HR-related document templates using AI.

For example, consider an employment contract when you hire someone. Instead of a general employment contract, there may be cases where highly specialized contracts must be concluded depending on the industry, business type, job type, or work involving dangerous tasks. In these instances, especially when hiring someone for the first time, you may not know what kind of contract to enter into.

In such cases, you can simply request the AI to create a template for the contract, saying, "I need to prepare a contract because this person is joining."

Let me show you two examples using the demo.

The first is an employment contract for an engineer.  If the contract involves working fully remotely from home, you can request, "Create a document that includes such conditions." The AI will then quickly generate and propose a template, saying, "How about this contract?" 

The second is creating a pledge at the time of onboarding. If you request to include conditions such as “Do not join a competing company for one year after resignation,” the AI will automatically generate and propose a template, saying, “How about this template?” 

The AI proposes document templates that are almost ready to use, and it can handle any level of complexity or special conditions. You can then conclude the contract electronically in the cloud and use it for employment contracts.

These are examples of how we utilize AI for auto-generation and proposals in the areas of accounting, contracts, and HR.

[New Initiative 1] AI Agent

Tsuji: As Hirohara just explained, AI features have been implemented in existing services, making them even more convenient.

In addition to this, we want to go beyond operational efficiency with AI and move to the next level.  Today, I'd like to highlight three key initiatives.

The first initiative is AI agents.

A well-known executive once said, "SaaS is dead,” and it became quite a topic in the industry. This slide illustrates what we envision AI agents to be.

Traditionally, with SaaS, users would perform actions such as inputting data or approving requests within the SaaS service. For example, clocking in on an attendance system, submitting an application on a year-end adjustment system, or checking credit card statements. These user actions triggered processing logic in the background, which then returned output to the user.

In the future, SaaS combined with AI will eliminate the need for users to manually input data in the UI, with AI agents completely replacing this process. Recently, voice input technology has also improved significantly, so both voice and text input will become possible. Therefore, we believe that AI agents will bring about significant changes in user experience (UX).

However, we think the underlying logic and data will not change much.  For example, in attendance management services, the logic and data related to clock-in/out will continue to run in the background. We believe logic and data will become even more important in the future.

Currently, our service areas include accounting, receivables, corporate performance management, spend management, contracts, HR, and SaaS management, and we believe we are probably No. 1 in the back-office product lineup.

By creating these back-office products, we have come to understand the best practices in each area.  We want to leverage this knowledge to build solid logic in the background and develop AI agents 

We believe there are three key points to the AI agents we envision.

• The first is to autonomously perform back-office tasks.  
• The second is to possess high expertise as a back-office specialist.  
• The third is to continuously learn and grow smarter from new data and logic.

Personalization is also a point we want to develop in the evolution of AI agents.

This is an image of the AI agent as we envision it.

Money Forward Cloud offers a variety of services, but we believe that our users utilize various tools, such as SaaS and core systems from other companies, communication tools like chat and email, and other document and spreadsheet software.

By integrating the data accumulated in the databases of these various business tools and connecting them all openly, we can create a data mart that contains a wide range of data across different tools.

We aim to create a system in which the data in the data mart is processed quickly and returned to the cloud, enabling AI agents to learn and perform tasks tailored to the corresponding business operations. This is the vision we are currently striving to realize.

This slide shows a simplified image of the data mart.

On the left side, there are Money Forward Cloud services such as Accounting, HR, and Contract. However, we not only incorporate data from our own services but also data from other companies' SaaS solutions and data updated by users themselves into the central integrated data mart.

The data mart holds modeled data.  While I might be stating the obvious, raw data is not usable. Therefore, we are currently creating a data mart that converts the data into an architecture or form that is easy for AI agents to use, organizes the data, and enables AI agents to learn appropriately.

Of course, security is also very important.  We are developing the data mart under the high level of security and governance required by internal audits.

Expense AI Agent

Tsuji: Let me introduce three specific services we are creating starting with the expense AI agent.

As many of you may have experienced, it can be quite a hassle to handle expense reports while busy with daily coverage, and it can be quite burdensome at the end of the month. It is quite a time-consuming task, but I think many people have to do it in a time crunch.

For example, if you use the Money Forward Business Card, the AI agent will detect the card use and suggest, “Your expense report for the card transaction is not done yet. Is that okay?” 

Furthermore, the AI agent will suggest, "If you just attach the receipt, I'll submit the report for you," and when you attach the receipt, the AI agent will infer the account category and department and submit the expense report as you can see in this diagram. 

In this way, the AI agent autonomously suggests actions, and users can simply respond with a “yes” button.

Then the approval request is sent to the approver in a flash, and you’re done with the application. I think you'll really feel like you have your own personal assistant.

Accounting AI Agent

Tsuji: The next service is the accounting AI agent.

Speaking of which, when I started my career, I was assigned to the accounting department right after graduating. Accounting work can be incredibly demanding, and there were times at the month-end and beginning when I couldn't go home due to overtime.

The reason I couldn't go home was often because the necessary data for processing wasn't sent from the field, which was quite stressful.  For example, expense reports weren't submitted, attendance records weren't submitted, or invoices weren't sent. These are common challenges accounting teams face. It was also mentally stressful to remind people, saying, "I'm sorry to bother you while you're busy with sales."

The AI agent handles all these tasks.

For example, the AI agent automatically identifies and reminds those who haven't submitted or approved.

Additionally, the accounting team will handle anomaly checks and close the accounts once all data is collected, but it takes time to gather all the data.  If possible, they would want to create a trial balance and perform anomaly checks even based on unapproved data.  AI agents can do that too.

The AI agent suggests, "I have created a list of unapproved expense reports and approvers for the provisional closing of the monthly accounts. Should I send reminders?"  

Then, it generates an estimated balance sheet, incorporating the currently unapproved data. It also highlights figures with significant variances from the previous month and includes a potential variance analysis. The values for the incorporated unapproved items are noted separately for clarity as well. It says if there's anything else, please make an additional request.

As you can see, the AI agent will be able to take over what accounting staff used to do.

After reviewing these suggestions and analyses, accounting staff can make a further request like, “Can you download the breakdown of travel and transportation expenses, including the unapproved items?” In response, a download file will be generated.  These capabilities will make accounting much easier.

So far, we have focused on an accounting team at small and medium-sized companies, but we are also planning to create various AI agents for accounting firms.

We plan to start from creating agents for document collection, sorting collected documents by type and client, recommending account titles, and creating monthly reports for clients.

HR AI Agents

Tsuji: The next service is the HR AI agent.

HR work is also very demanding, with a wide range of tasks. Lately, human capital disclosure has been particularly burdensome. Data is scattered in various places, and sometimes the data that comes out is wrong. It's a major pain point for us too.

For example, as data related to human capital, you can request the AI agent to provide parental leave acquisition rates by department and job type.  The AI agent will then search for the data and provide an answer.

Moving on to the next slide, which specifically asks the AI agent, in preparation for human capital disclosure, to compile data on male parental leave acquisition rates by department and job type, graph it in an easy-to-understand manner, and then comment on what trends you see.

The AI agent will then search for the data and create a graph like this to visualize the results. It will also provide comments on overall trends, departments with high acquisition rates, and future implications.

Furthermore, anticipating that your boss will likely ask, "How do we compare to other companies?," if you request a comparison with them, the AI agent will add the average values of listed companies to the graph, summarizing the differences in acquisition rates compared to our company's.

What used to require us to acquire and compile data ourselves, piece by piece, can now be done easily in this way.  We will continue to create such AI agents for our various services and make them even more convenient.

[New Initiative 2] AI Agent Platform

Tsuji: The second initiative is that we want to move forward with the AI agent platform concept.

We will, of course, aggressively develop our own AI agents. However, there will be areas where we fall short or things we can't do ourselves only, and more convenient AI agents will continue to emerge in the world.

Therefore, we envision that by opening up our cloud services and seamlessly integrating with agents from our partners and other companies through APIs, we can create even greater convenience for our users. We want to take on the challenge of bringing this worldview to life.

This is similar to the App stores on smartphones. We have a large user base, so partners who can join our platform will be able to take advantage of our user base, and we will, of course, compensate them for the revenue generated there. We are considering charging a platform fee, but the details are not yet decided.

In this way, partners who create good AI agents can acquire more users, grow their business, and our users' lives will become more convenient. And that’s the worldview we want to create.

We therefore envision a wide variety of agents emerging, such as expense reimbursement agents or invoice processing agents, and we believe that is an exciting aspect of our worldview.

We have always valued the concept of “openness,” and we aim to evolve back-office operations and provide an open and integratable AI agent platform to realize an ecosystem.

Therefore our aspiration is to create an ecosystem and worldview that connects us with various companies to deliver more convenient services to our users. 

[New Initiative 3] AX Consulting

Tsuji: As our third initiative, we are announcing the start of AX consulting. 

We have already been providing quite a lot of business consulting services for back-office operations, but we have decided to leverage AI more and expand our services to mid-sized and enterprise companies, not just SMEs.

Currently, we are promoting back-office consulting for over 500 companies with three companies: Money Forward Cloud Corporate Performance Management Consulting, Inc., Knowledge Labo, Inc., which once was our customer but joined our group, and OutlookConsulting Co., Ltd., which became part of our Group through a recent TOB.

Based on the consulting experience of these three companies, we aim to further promote AX.

Let me then talk about the service content of our AX consulting.

Recently, a friend of mine who is the president of a large company in a regional area consulted me, saying, "I want to improve the back office, but I have no idea how. I can't hire people, and I don't have specialized employees. What should I do?"    

In such cases, we first review the entire back-office operations, break down the tasks, and consult and propose which services would be best suited for each operation.

Without in-depth consulting that suggests, for example, "Considering your current operations, transitioning to this service would be beneficial," it's challenging to improve the back office.

Therefore, we first identify the challenges, design the optimal business processes, and then propose, "These SaaS and AI agents will make your work easier and more appropriate."  What's crucial in fact is to ensure that the new processes actually work once implemented, so we plan to develop a service that comprehensively provides operational support.

Additionally, our Group company Money Forward X, Inc. currently provides services to over 80 financial institutions, and we plan to expand our AX consulting services to the financial institutions as well.

Financial institutions also have a lot of paper-based and manual input work in their back-office operations, so we plan to expand consulting for them as well.  This would be more of a comprehensive AX consulting service rather than just back-office consulting.

Conclusion

Tsuji: We aspire to become the No.1 back-office AI company.

Firstly, we will introduce AI agents into all our services to automate and autonomize back-office operations. 

Secondly, we will provide an open AI agent platform to build an ecosystem of AI agents for back-office operations. 

Thirdly, we will start offering AX consulting services to mid-sized and enterprise companies, as well as financial institutions.

These are the three initiatives we plan to pursue.

This will be our timeline. Firstly, regarding AI agents, we will progressively roll them out throughout 2025

Second, regarding the AI agent platform, we will now start accepting applications from our partners who wish to work with us from today. We hope you will reach out to us and help create an open worldview together.

Third, for AX consulting as well, we will start accepting inquiries from today.  We hope to be able to help those struggling with back-office issues.

I personally believe that the AI impact Money Forward aims for has even greater potential.

We aim to more than double productivity and efficiency across operations, sales, and development. Our management goal is also to more than double sales per employee. Furthermore, we see significant potential to enhance the value we provide through new features and products leveraging AI, and we plan to actively pursue this as well.

As technology evolves, so will AI continue to be updated. We will update our AI vision accordingly and strive to deliver better services to our users. Thank you.