The Era of 8,300 Partners Using AI by Voice
In July 2025, the GMO Internet Group launched a new project. Called “Project Whisper for AI by GMO,” this initiative provides approximately 8,300 partners across the group with the ability to “use AI by voice.”
At first glance, this may seem like just another internal tool rollout. But for business leaders, it should be seen as a major turning point in the “democratization of AI adoption.”
Why? Because the biggest barrier to AI tool adoption has always been “operational complexity.” For those who struggle with keyboard input or are unfamiliar with PC operations—like field staff—text-based AIs like ChatGPT and Claude have felt out of reach. Voice input has the potential to tear down that wall in one go.
Three Barriers Voice AI Breaks Down in the Workplace
I personally use Claude Code and ChatGPT in my daily work, and I’ve noticed that the quality of thinking changes between text and voice input. Typing often interrupts my train of thought, but with voice, I can toss ideas directly from my mind to the AI.
For management, three key points stand out:
First: Speed
Voice input is said to be about 3 to 4 times faster than typing. Tasks like meeting minutes, brainstorming ideas, and drafting replies to clients can be dramatically shortened.
Second: Low Barrier to Entry
Even veteran employees who aren’t comfortable with PCs, or staff focused on hands-on work, can use AI by simply pressing the microphone button on their smartphone and speaking. This low barrier is exactly why GMO can roll it out to 8,300 partners at once.
Third: Compatibility with Multitasking
Even when your hands are busy, you can give instructions to AI using just your voice. On a factory line, during sales calls on the go, or while cooking in a restaurant—voice AI can be introduced to workplaces where AI adoption was previously difficult.
Realistically Considering Implementation Costs and Hurdles
As a business leader, cost is likely a concern. While GMO’s case is geared toward large enterprises, small and medium-sized businesses can easily adopt similar systems.
Here are some specific options:
1. Use Voice Input Features in Existing AI Tools
The ChatGPT app comes standard with voice input. It’s available on the Plus plan for $20 per month (around ¥3,000). The Claude app offers the same. This is the lowest-cost way to get started.
2. Task-Specific Voice AI Services
Specialized services are growing, like “Notta” and “AI GIJIROKU” for meeting minutes, or “Salesforce Einstein” for sales support. These can be implemented for around ¥5,000 to ¥20,000 per month.
3. In-House Development Options
By using speech recognition APIs (like Google Speech-to-Text or Whisper), you can build a voice AI system optimized for your own business. Initial development costs range from hundreds of thousands to millions of yen, but in the long run, it offers the advantage of breaking free from SaaS dependency.
Three Points to Watch When Introducing Voice AI
However, voice AI also has challenges. Here are the key points for leaders to consider before implementation:
Privacy and Security
Voice data can contain even more sensitive information than text. In GMO’s case, careful attention is likely paid to how voice data is handled. Establishing internal rules and verifying how the service you use handles data are essential.
Specifically, check the following:
- Whether voice data is used for training
- Data retention periods and deletion policies
- Measures for handling information that cannot be taken outside the company
Speech Recognition Accuracy
Japanese speech recognition accuracy improves every year, but it can still struggle with technical terms, proper nouns, or dialects. Before implementation, I recommend testing the terms used in your own business.
In my case, legal terms and contract-specific jargon weren’t recognized accurately at first, causing some trouble. As a countermeasure, choose a service with a custom dictionary feature, or set up a system for regular feedback.
Organizational Culture and Practices
Giving instructions to AI by voice is still not common. Especially in Japan, many employees may feel reluctant to “work out loud.” Before introducing it, it’s important to share rules and etiquette for usage within the company and adopt a strategy for gradual adoption.
The Management Impact of Voice AI
GMO’s case isn’t just about introducing a tool. Equipping 8,300 partners with the ability to use AI by voice is an effort to raise the organization’s overall “AI literacy.”
As a business leader, the question to ask is this:
“Are our employees blocked by the ‘operational wall’ of using AI?”
If AI adoption is limited to a few highly IT-literate employees, voice input is an effective way to tear down that wall.
In my experience, the impact of AI adoption scales with the number of users. Having 100 regular users each save one hour per month contributes more to overall productivity than one super-user saving 100 hours per month.
Voice AI is a tool that accelerates the latter—spreading AI to the average user.
Summary: What’s Your Next Move?
GMO’s “Project Whisper” signals a new phase in AI adoption: a shift from “people who can use AI” to “anyone can use AI.”
Here are three actions you can take right now as a business leader:
- Try it yourself first: Use the voice input feature on the ChatGPT or Claude app in your daily work for just one week. Your own experience is the best basis for judgment.
- Consider a pilot rollout: Test voice AI tools in a specific department or team and measure the results. Starting with routine tasks like meeting minutes or daily reports is a safe bet.
- Establish internal rules: Create guidelines that clearly define how voice data is handled, the scope of use, and prohibited actions. Balancing security and convenience is key.
Voice AI is still a nascent technology. But the fact that major companies like GMO are actively adopting it is proof that this trend is real. Experiencing it now and identifying the right implementation method for your company will be key to building a competitive edge.
Comments