The Ultimate Solution for “Breaking SaaS Dependence” Revealed by Japan’s New AI Consortium

The New Domestic AI Company Represents More Than “Tech Development”—It’s a Management Innovation

SoftBank, NEC, Sony, and Honda have taken the lead in establishing a new company, “Japan AI Infrastructure Model Development.” Much of the media coverage focuses on the technical aspect of “breaking dependence on foreign AI.” However, from a management perspective, this holds deeper significance. It marks the beginning of the final stage for Japanese companies to break free from the structural challenge of “SaaS dependence” and to in-house their core competitive “knowledge foundation.”

In my own company, I have concurrently used Claude, ChatGPT, and Grok, achieving a reduction of 1,550 hours annually across 93 business areas. Through this process, I acutely felt the “data siloing” and “hollowing out of competitiveness” brought about by reliance on foreign SaaS. This recent move can be seen as a definitive response from Japanese companies to this challenge.

The “AI Cultivation Without Copy-Paste” Indicates the Next Stage

The simultaneously reported keyword “AI cultivation without copy-paste” is telling. It points out that the “next correct answer” for generative AI adoption in large corporations lies not merely in introducing tools, but in embedding the capability to “cultivate AI” within the organization. This precisely signifies breaking away from SaaS dependence.

SaaS is convenient, but its capacity for customization with a company’s own data is limited. Furthermore, relying on external services for critical business logic and know-how poses a risk from a long-term competitiveness standpoint. This domestic AI development initiative holds the potential to fundamentally change this dependency structure.

Three “SaaS Dependence” Risks That Managers Overlook

Many managers focus on the convenience of SaaS—getting the latest features for a fixed monthly fee. However, they tend to underestimate the following three risks.

Decision-Making Delays Due to Data Silos

Introducing multiple SaaS platforms like Salesforce, Slack, and Notion traps data within each service. This is “data siloing.” For instance, sales data, customer communications, and project information become scattered across separate platforms. This makes integrated analysis difficult and slows down decision-making.

Leakage of Competitive Advantage Sources to External Parties

When unique business processes or know-how are incorporated into a SaaS workflow, the very structure may become known to the service provider. Especially when utilizing AI features, there’s a risk that insights into “how your company achieves results” could indirectly become a platform leveraged by competitors as well.

Cost Structure Rigidity and Risk of Sudden Price Hikes

SaaS usage fees become recurring expenses. As the market matures and alternative services dwindle, price increases can occur. In fact, price revisions for major cloud services are not uncommon. If core business operations depend on a specific SaaS, your company’s bargaining position in price negotiations weakens.

The “In-Housing” Practical Solution Enabled by the Domestic AI Infrastructure

The foundational model to be provided by this newly established company could offer a practical solution to these risks. Specifically, the following management benefits are conceivable.

True Customization with Company Data

With generic overseas AI models, you cannot train them using your company’s confidential data like past contracts, customer service histories, or technical notes. However, if you can operate a domestic foundational model within your own environment, it becomes possible to develop specialized AI that leverages this valuable data. This goes beyond mere operational efficiency to become a source of competitive advantage.

Automation of Data Integration Across Multiple SaaS Platforms

By maintaining your own AI infrastructure, you can use it as “glue” to extract and integrate data from various SaaS platforms. For example, you could build a system that integrates customer information from Salesforce, communications from Slack, and sales data from your internal database, with AI proposing the next sales action to take. This can resolve data siloing.

Improved Cost Predictability

When operating an AI model on your own servers or domestic cloud, you can design a cost structure based on your own usage volume. This frees you from the risk of sudden price hikes and makes long-term IT investment planning easier.

Three Steps for SMEs to Start Preparing for “In-Housing” Now

The moves by the large corporate consortium are on a massive scale and cannot be immediately replicated. However, even managers at small and medium-sized enterprises (SMEs) can start preparing for in-housing now. Begin with the following three steps.

Step 1: Digitize and Structure Your Company’s “Knowledge Assets”

First, identify the know-how that forms the source of your company’s competitiveness and digitize it. Examples include negotiation records from top salespeople, troubleshooting logs from engineers, and highly praised customer service cases. Accumulate these not as mere notes, but as structured data (e.g., in Q&A format, case study format). This task can also be automated using ChatGPT or Claude to extract information from existing documents.

Step 2: Begin Internal Experiments with Open-Source AI Models

There’s no need to start by operating a large-scale model immediately. First, try running lightweight open-source AI models like Llama 3 or Gemma on internal PCs or a small server. Costs can start from a few hundred dollars per month. This allows you to experience the behavior when feeding it your company’s data and the required computational resources. This experimental phase will also reveal the personnel and skill requirements needed for future in-housing.

Step 3: Identify Business Processes with High SaaS Dependence

Audit which current business processes depend on which SaaS platforms. Then, assess the business impact (BIA) if that process were to stop. The greater the impact, the higher the priority for in-housing or multi-vendor strategies. This analysis forms the foundational data for strategic IT budget allocation.

Concrete Action Plan for Managers

In light of this news, here are actions managers can tackle this week.

1. Schedule a Dialogue with Your CTO or IT Lead

Pose these questions: “To what extent do our critical business processes depend on which SaaS platforms?” “What is required to extract data from these SaaS platforms and perform integrated analysis?” Discuss this from the perspective of business risk and opportunity, not technical details.

2. Launch One Small In-Housing Project

For example, try in-housing the automatic generation of your monthly sales report using open-source AI. This is an experiment to build a pipeline that generates reports directly from your company’s data, instead of using a SaaS reporting feature. Start with a small project: budget under ~$3,000 USD and a timeline under three months.

3. Establish a System to Monitor Developments in the Domestic AI Infrastructure

Gather information on when and under what conditions the services offered by the new company “Japan AI Infrastructure Model Development” will become available. Through your industry association or business partners, explore possibilities for early access or pilot program participation. Don’t wait passively for information; proactively utilize your network.

The Future is a Hybrid of “Owned AI” and “Rented AI”

In conclusion, future management strategy is not a binary choice between in-housing everything or depending entirely on SaaS. What’s important is a “hybrid strategy.”

For the “knowledge foundation” related to your company’s core competitiveness, aim to in-house (own) it as much as possible, utilizing domestic infrastructure. On the other hand, continue using (renting) SaaS for generic business applications (email, calendar, expense management, etc.). Clearly defining this division is the new responsibility of management.

The significance of giants from different industries—SoftBank, NEC, Sony, Honda—joining forces is profound. This is not merely a technology development project; it is an attempt to change the very nature of the digital infrastructure for the entire Japanese economy. How each manager incorporates this major trend into their company’s strategy will be decisive. That thinking and action will likely determine corporate destinies over the next 5 to 10 years.

Start by creating a “SaaS Dependence Map” for your company. There, you should find the compass for your future investments and in-housing efforts.