Conclusion

Let’s get straight to the point. “Saving prompts and executing them with one click” has become a standard feature in the browser.

Google has added a new feature called “Skills” to Chrome. What it does is simple: it allows users to save prompts for AI and call them up with one click on any webpage. That’s it.

However, the implications of “that’s it” are significant.

Until now, when trying to use AI for business, users faced three barriers: “having to retype prompts every time,” “switching tools,” and “not knowing what to ask in the first place.” Skills eliminates the first two barriers, and it’s free.

I want to ask: Is the monthly fee you’re paying for business efficiency tools at your company really justified?

—

What Happened — Understanding Skills in 30 Seconds

The structure of the feature is as follows:

1. Enter a prompt into Gemini (Google’s LLM) next to the Chrome address bar.
2. Select “Save this prompt as a Skill.”
3. From then on, regardless of which webpage you have open, you can execute the saved Skill with one click.

For example, if you save a prompt like “Summarize the content of this page in 300 characters and create a report email for the client,” you can run the same operation with one click whether you’re researching a competitor’s site or checking industry news.

The key point is that it is completed within the browser. There’s no need to launch another app, nor to copy and paste text. You can pass the content of the open page directly to the AI.

—

Let’s Talk Costs — This is the Main Issue for Small and Medium Enterprises.

When local small and medium enterprises try to use AI in their operations, they generally face these options:

If a company with five employees uses ChatGPT Team, it would cost 22,500 yen per month, totaling 270,000 yen annually. Adding just one specialized tool could bring the total to between 400,000 and 800,000 yen per year. For small and medium enterprises, this amount is not insignificant.

While the range of what Skills can cover is limited, there are certainly cases where “repeating standard AI processes while viewing a webpage” is sufficient.

Let’s be specific.

—

Three Practical Use Cases

1. Automated Summarization for Competitive Research

Imagine a sales representative who checks the news release pages of five competitors every morning. Traditionally, they would read each page and compile the information into Excel, taking about 15 minutes per company and 75 minutes total.

If they save a prompt in Skills like “Classify the new information on this page into bullet points: ① Product updates ② Price changes ③ Campaigns,” they can execute it with one click on each page. The task of reading and summarizing becomes just a matter of verification.

75 minutes can be reduced to 15 minutes. Calculating for 20 business days a month, that’s a reduction of 20 hours. At an hourly rate of 2,000 yen, that’s 40,000 yen a month, or 480,000 yen a year in labor costs.

2. Draft Generation for Inquiry Emails

When replying to emails from clients, the content is often repetitive. Common inquiries include “delivery date confirmation,” “estimate requests,” and “initial responses to complaints”—typically falling into 5 to 10 patterns.

Previously, users would copy and paste template responses and manually rewrite them. With Skills, they can save a prompt like “Read the content of this email and generate a reply based on the following rules.” When they execute the Skill while opening Gmail, a draft that fits the context of that email will appear.

If the response time per email shrinks from 10 minutes to 3 minutes, handling 20 emails a day saves 140 minutes. That’s 46 hours a month, equivalent to about 1,100,000 yen annually (at an hourly rate of 2,000 yen).

Of course, final verification and sending will still be done by a person. However, the mental load of “writing from scratch” versus “reviewing and correcting” is vastly different.

3. Information Gathering for Recruitment Pages

Recruitment is always a challenge for local small and medium enterprises. They need to understand not only their own job postings but also the conditions under which competitors are hiring.

By creating a Skill that says, “Extract job type, salary, location, and benefits from the job information on this page and put it into a table,” they can obtain structured data with one click every time they open a competitor’s recruitment page.

What used to take the HR department half a day to compare competitor job postings can now be completed in 30 minutes.

—

The Real Importance Is Not Just ‘Cost Reduction’

While we’ve talked about costs, let’s be honest: the essential value of Skills is not merely “cost reduction.”

It’s about creating a system that eliminates the gap between those who can use AI and those who cannot.

The primary reason AI utilization is slow in small and medium enterprises is that there are only one or two people in the company who can write prompts. If that person leaves, it’s over. It’s the epitome of dependency.

Skills transforms prompts into “skills that can be saved and shared.” If there’s one person who can write effective prompts, everyone can use that Skill. While it may sound grand to say it democratizes prompt engineering, essentially it means “the know-how of capable individuals can be distributed to everyone with one click.”

This is more effective for small and medium enterprises than for large corporations. Why?

Large companies have IT departments that can develop specialized tools. Small and medium enterprises do not have that. Therefore, the benefits of a “free systematization tool that is completed within the browser” are substantial. In smaller organizations, the speed at which one person’s know-how can be disseminated to everyone is faster. In a company of five, everyone can start using the same Skill from tomorrow.

—

Calmly Assessing Risks and Limitations

It would be irresponsible to only praise Skills without acknowledging its current limitations.

1. The information processed is primarily from “the currently open page.”
Cross-page analysis or integration with internal databases is not possible. The main battlefield remains text processing within the browser.

2. Handling of confidential information
Passing page content to Gemini means there’s a possibility that data is sent to Google’s servers. It’s dangerous to transmit customer personal information or contract details directly. Establishing internal rules is necessary beforehand.

3. The accuracy of output depends on the “quality of the prompt.”
If the prompts saved as Skills are poorly constructed, everyone will end up using the same low-quality output. “Who designs the Skill” merely shifts the issue of dependency to another layer. A mechanism for regular review is needed.

4. Currently available only as a Canary version (developer test version)
The timeline for official inclusion in Chrome for general users is undecided. It’s not something you can use today. However, the fact that Google has steered in this direction is significant.

—

So, What Should We Do?

Here are three actions to take:

① Right now: Take inventory of your company’s “repetitive prompts.”
Ask those in your company who use AI, “Aren’t you typing the same prompts every time?” That list will become the “initial skill set” when implementing Skills.

② Soon: Try the Chrome Canary version.
If you have one tech-savvy employee, they can install the Canary version and give it a try. There’s no need to wait for the official version. Companies that engage with it first will gain a head start in know-how.

③ Concurrently: Establish rules for handling information.
Decide in advance what information can and cannot be passed to Skills. It’s too late to rush and decide after the official release.

—

An Era Where Browsers Become ‘Business Tools’

Looking back, Excel became a business platform from a spreadsheet software, and email transitioned from a communication tool to the core of business workflows because they were both things that were “opened every day without fail.”

The browser is currently the application that is most frequently opened by small and medium enterprises. AI’s standardized processing is now available for free within it. I believe this marks the beginning of a tectonic shift in business tools.

I’m not saying that the 30,000 yen monthly business efficiency tool will become unnecessary. However, it’s essential to consider how much of that 30,000 yen can be replaced by Chrome now.

What happens when costs decrease? The answer is clear: “the reasons not to do it disappear.” Even a small company with ten employees can access AI utilization at zero monthly cost. The only question left is whether to act or not.