The Quiet Revolution in Finance
News has been buzzing about Sumitomo Mitsui Financial Group (SMFG) collaborating with AI startup Sakana AI to develop a system that creates strategic proposals for client companies in just hours.
Traditionally, when banks create strategic proposals for clients, industry analysis, financial analysis, and market research typically take weeks or even months. However, with this new system, AI instantly analyzes vast amounts of data, allowing human experts to review and finalize proposals in a fraction of the time.
This isn’t just about “operational efficiency.” It’s a paradigm shift in the financial industry that changes the very speed of business decision-making. This article explores the essence of AI adoption that business leaders and CTOs should learn, and how it can be applied to their own organizations.
Why Fintech AI is Gaining Attention Now
The financial industry has always been proactive in data analysis, but the emergence of generative AI has completely changed the landscape. Traditional AI primarily focused on “learning patterns from past data and making predictions.” In contrast, generative AI can “create new value.”
In the SMFG case, by leveraging Sakana AI’s technology, the AI generates specific business strategies and even M&A proposals based on client financial data and market trends. Human analysts shift to a role of verifying and refining the AI’s output.
This shift reveals that AI is evolving from a “analysis tool” to a “strategic partner.” For business leaders, this trend isn’t limited to the financial industry. Similar transformations could occur in their own strategic planning processes.
How Faster Strategic Proposals Transform Management
Traditional strategic proposal processes faced several challenges:
- Time-consuming data collection
- Dependence on the skills of individual analysts
- High effort required to compare multiple options
- Inability to keep pace with changing business environments
What changes when AI solves these challenges?
First, the decision-making cycle is dramatically shortened. SMFG’s system reportedly enables proposals to be created in “just hours.” This allows executives to make decisions “today” instead of “by the end of the week.”
Second, the number of options that can be considered increases. Where a human analyst might take a week to evaluate three options, AI can generate 30 options in half a day. The executive’s role shifts to selecting the best option from a wider range.
Third, the risk of knowledge silos is reduced. Even if a veteran analyst leaves, the knowledge accumulated by the AI remains. This has a similar effect to the process I practice in my own company: “standardizing tasks → anonymizing them → storing them in a shared drive.”
3 Steps to Apply This to Your Company
While the SMFG case might seem geared towards large corporations, there are points applicable to small and medium-sized businesses. In fact, I’ve seen similar approaches yield results at my consulting clients.
Step 1: Organize Data Usable for Strategic Planning
To have AI propose strategies, you first need to organize your company’s data. Structure the information you’ll feed the AI: financial data, customer data, market data, competitor data, etc.
Many companies stumble here. Their data is scattered across Excel files and PDFs, not in a unified format. The first step is to think about creating a “database for AI to read.”
Specifically, it’s helpful to compile data like:
- Financial statements for the past 3 years
- Key customer attribute data
- Public information on competitors
- Industry market reports
Step 2: Give AI a “Thinking Framework” Through Prompt Design
When asking AI to propose strategies, the most important thing is the prompt design. Vague instructions like “Think of a strategy” won’t yield high-quality output.
For example, giving instructions within a framework like this is effective:
“You are a management consultant. Analyze the following data and propose three business strategy options. For each option, clearly state the pros, cons, required resources, and estimated ROI. Also, prioritize them based on feasibility.”
By clearly instructing the AI on its “role” and the “format of the output,” you can get proposals that are practical enough for real-world use.
Step 3: Verify and Refine with Human Judgment
AI-generated strategy proposals are just “drafts.” The final decision must always be made by a human. Even SMFG’s system includes a process where experts verify the AI-generated proposals.
In my experience, AI proposals are at an “80-point” level. The remaining 20 points are completed by incorporating industry-specific nuances and factors not captured in the data, like human relationships.
If you can establish this process of “AI creates the draft, humans refine it,” the speed of strategic planning can increase by more than three times.
Implementation Costs and Hurdles
Building a large-scale system like SMFG’s in-house isn’t realistic for most, but there are more accessible methods.
For example, paid plans for ChatGPT or Claude (around $20-$30 per month) can generate sufficiently practical strategic proposals with proper prompt design and data preparation. I personally use Claude for creating client proposals, generating value equivalent to approximately $50,000 annually for a monthly cost of about $140.
For a more serious approach, consider tools like:
- Sakana AI: A generative AI service specialized in finance and business strategy
- Notion AI: Enables strategic planning based on internal knowledge data
- Custom GPT: Building a dedicated AI trained on your company’s data
The hurdles to implementation are less about technology and more about operational aspects: “How do we organize our data?” and “How do we effectively use AI?” I recommend starting with a small project and building on your successes.
What Business Leaders Should Do Now
The SMFG case offers valuable insights for leaders in all industries, not just finance. The automation of strategic proposals using AI is no longer a “future concept” but a “current reality.”
Here are three things business leaders should tackle immediately:
First, take stock of your company’s data assets. Understand what you can input into AI and what’s missing.
Second, actually try AI tools. Start with free trials or low-cost plans and test how much they can be used in your business operations.
Third, drive a cultural shift within your organization. Foster a positive attitude that focuses on “collaborating with AI to do higher-value work” rather than the fear of “jobs being taken by AI.”
AI evolution is relentless. Whether you act now will determine your competitive edge in the years to come.
Comments