- What the End of ChatGPT’s Dominance Means
- The Essence of the Decline: Changing User Demands
- The Common Trait of Companies That Turn “Brilliant AI into Incompetent Employees”
- Three Actions Leaders Must Take Now
- Cost and Implementation Hurdles: Realistic Decision Criteria
- Summary: Choosing an AI Is a Management Decision in Itself
What the End of ChatGPT’s Dominance Means
A major tectonic shift is underway in the generative AI market. According to reports from Business + IT, ChatGPT’s market share has fallen below 50% for the first time, ushering in a “Generative AI Tri-Power Era” where Google’s Gemini and Anthropic’s Claude are rapidly closing the gap.
For executives and CTOs, this news is more than just a trend shift. It’s a “turning point” that forces a fundamental reassessment of your company’s AI investment strategy.
Why? Because companies that previously defaulted to “just use ChatGPT” will now face the complex challenge of deciding “which AI to use for which task, and how to combine them.”
The Essence of the Decline: Changing User Demands
ChatGPT’s market share drop reflects a “maturing” user base. The initial excitement of “I just want to try generative AI” has cooled, and we’ve entered a phase where users are selecting the optimal tool for each specific task.
In fact, in my own work, the division of labor is clear. I use Claude for contract reviews and complex legal analysis, ChatGPT for creative copywriting, and Grok for real-time information gathering—using three different AIs depending on the purpose.
This “multi-AI strategy” is precisely the mindset required for future management.
The Common Trait of Companies That Turn “Brilliant AI into Incompetent Employees”
An article from Keyman’s Net presents shocking data: fewer than 20% of companies that introduced AI achieved “increased sales.”
Why does brilliant AI end up becoming an “incompetent employee”? Based on my consulting experience, the causes boil down to three factors.
Cause 1: Expecting AI to Be a “Jack of All Trades”
Telling ChatGPT to “increase sales” won’t yield specific actions. AI is merely a tool that replaces “parts of a task”—the overall strategy must be defined by management.
Cause 2: Over-Reliance on a Single AI
Continuing to use only ChatGPT while ignoring the unique strengths of Gemini and Claude is like giving every employee the same hammer. This approach lacks the perspective needed to select the best AI for each task.
Cause 3: Not Knowing How to “Nurture” AI After Implementation
AI implementation is not the finish line. Continuous tuning is required—improving prompts, adding training data, integrating into workflows. Neglecting this turns AI into a “useless tool.”
Three Actions Leaders Must Take Now
In the Generative AI Tri-Power Era, the strategy for leaders is clear. Please implement the following three actions.
Action 1: Adopt a Multi-AI Strategy
At a minimum, introduce ChatGPT, Gemini, and Claude, and understand the strengths of each. Based on my results, the total monthly cost is approximately 21,000 yen (about $140). This enables the creation of value equivalent to roughly 7,533,000 yen (about $50,000) per year.
Specifically, the following division of labor is effective:
- ChatGPT: Marketing copy, customer email creation
- Claude: Contract review, internal policy analysis, code generation
- Gemini: Google Workspace integration, data analysis
Action 2: Clearly Define the “End Goal” of AI Implementation
Instead of “introducing AI,” define “what you want to achieve with AI.” For example, set a concrete KPI like “reduce contract review time from 20 hours per month to 5 hours.” This makes it possible to measure AI’s effectiveness.
Action 3: Raise AI Literacy Across the Company
It’s meaningless if only management can use AI effectively. Implement training and study sessions so all employees can decide “which AI to use for which task.” In particular, everyone should master the basics of prompt engineering.
Cost and Implementation Hurdles: Realistic Decision Criteria
The arrival of the “Generative AI Tri-Power Era” is also good news from a cost perspective. As competition intensifies, companies will engage in price wars and feature enhancements. In fact, Claude now offers high-quality analysis even on its free plan, and Gemini adds value through integration with Google Workspace.
In terms of implementation hurdles, most AIs can be started for a few thousand yen (tens of dollars) per month. The key is the “design philosophy” of which AI to use for which task and how to combine them—the initial investment is extremely low.
However, there are points to watch out for. Over-reliance on a single AI creates the risk of vendor lock-in (a state where switching becomes difficult due to dependence on a specific AI). A multi-AI strategy also serves as a hedge against this risk.
Summary: Choosing an AI Is a Management Decision in Itself
ChatGPT’s market share decline is evidence that the generative AI market has moved from the “experimental stage” to the “practical stage.” What is required of leaders is the attitude to choose and nurture the AI best suited to their company’s tasks, without being swayed by trends.
My own multi-AI strategy, which achieved a reduction of 1,550 hours per year, is by no means special. Anyone can start with an investment of around 20,000 yen (about $130) per month. What matters is “starting” and “continuing to nurture.”
In the Generative AI Tri-Power Era, which AI will your company leverage as a “management resource”? That decision will determine your competitive edge.
Comments