A Black Ship has arrived at the call center. The day an SV with 7 years left until retirement realized they couldn't escape.
SoftBank partners with Sierra. Exclusive distribution of conversational AI platform in Japan
The moment I saw that headline, I couldn't move from in front of my screen.
For me, a 58-year-old active call center SV, this wasn't just any AI news.
Finally, the premises of our work are changing.
The moment I thought that, a chill ran down my spine, and a strange sense of impatience welled up.
I, who am supposed to be the 'AI promotion lead' on the floor, calling on members every day to use AI, was the one most afraid after seeing that news.
Seven years left until retirement.
It's not that I thought I could escape it. But I felt as if a cold reality had been thrust upon me, that the countdown to that 'end' had finally begun in earnest.
A 97% resolution rate. This is a number that means not 'fewer calls,' but 'seats disappearing.'
As I read through SoftBank's official release, numbers that made me doubt my eyes were lined up.
It stated that as a result of a pilot test on the online-only brand 'LINEMO,' the resolution rate for inquiries improved from 83% to 97%.
Anyone who has worked in a call center should know how extraordinary this number is.
Conventional chatbots were, so to speak, 'machines that just distribute answers.' They would search prepared FAQs and just provide a link saying, 'Please read this.' So in the end, customers couldn't resolve their issues and would call, saying, 'I don't understand, so I want to talk to someone directly.' 'I used the chat, but the calls haven't decreased at all.' Such complaints were a daily occurrence on the floor.
However, this 97% figure is different. It means that the AI completed the 'procedure the customer really wanted to do' right there within the chat. SoftBank's presentation materials stated that Sierra is a 'goal-oriented AI agent.' I want to return an item, I want to cancel, I want to change my plan. For such objectives, the AI judges the situation itself and carries out the necessary procedures. It doesn't just hand over an answer and finish; it executes the work itself to the very end.
In my workplace, there is a veteran operator I rely on most. A while ago, when I talked to them about AI, they said this:
'AI is difficult. I don't feel like I can use it. I feel like I'm being told I'm no longer needed.'
At that time, I could only offer words of comfort.
But now, seeing the 97% figure, the weight of those words has returned with a different meaning. The 'processing of procedures' and 'confirmation exchanges' that make up much of the work we do every day—managing about 30 operators and coordinating with about 60 clients—is a world where those tasks are completed to a high degree of accuracy without any human intervention, and it already exists on the other side of the screen.
In other words, this is not just a story about the number of calls decreasing. It is a story about the very seats prepared to receive those calls becoming unnecessary.
The Japanese contact center market is screaming due to a serious labor shortage and soaring labor costs. That is precisely why large companies are seeking this technology. While I, a mere SV with no budget or authority, am treading water—shouting 'AI promotion' while being blocked by my company's security walls and unable to input even a single character of customer information into AI—mega-carriers are building 'call centers that do not rely on human labor' at overwhelming speed.
Sierra is not just an extension of playing around with ChatGPT.
To summarize this company called Sierra in one sentence, it is this.
It is not a 'convenient chatbot,' but an AI agent that major global corporations and leading banks are already entrusting with their actual operations.
Co-founder Bret Taylor is a former OpenAI board chair and former Salesforce co-CEO. The other, Clay Bavor, also served as a leader at Google. These two did not create an AI for general consumers to play with, but an AI designed to carry a company's brand and operate in real-world production environments. That is precisely why security considerations, such as not using customer data to train models, are built in to a significant degree. Moreover, it is integrated not just with text chat, but also with voice. The future where AI autonomously handles phone support itself is already right before our eyes.
The AI I thought was an overseas story has arrived at the front lines in Japan.
However, what I was truly afraid of was not the performance of Sierra itself.
It was that the changes I thought were an overseas story have officially landed on the front lines in Japan. That was what I feared most.
I have written about Sierra here twice before. Once about performance-based AI agents, and another time about an AI called 'Fin' from Intercom. Both were written after I was shocked by reading articles written by Ryohei Katayama, the founder of paiza.
And thankfully, thanks to those two articles, I even had the opportunity to speak with Mr. Katayama one-on-one online for about an hour.
Mr. Katayama's insights, being well-versed in overseas affairs, were undoubtedly valuable. But to be honest, a feeling close to 'someone else's story' still remained somewhere within me.
I had a premonition that this kind of technology would eventually come to Japan from overseas. But I thought it would be a little further in the future.
I have also learned a lot from the posts of 'Ope-Suisui,' who shares information about overseas call center situations on X (formerly Twitter). Even so, there was a part of me that was taking it easy, thinking, 'It will still take time for these kinds of services to fully penetrate Japan.'
If I were to honestly put my impression of seeing SoftBank's announcement this time into words, it would be this.
Finally, a Black Ship has appeared off the coast of Uraga.
The shogunate officials at the end of the Edo period must have naturally received information that 'the Black Ships are coming, they are coming.' Even so, their sense of crisis as a reality was thin. It's not that they were doing nothing. Even so, the shock at the moment the ships actually appeared must have been something entirely different from the amount of information they had beforehand.
It is probably the same for me and the people in the call center and contact center industry right now. We knew in our heads that 'AI will eventually come.' Even so, the shock of the moment I saw the sentence 'Exclusive sales agent in Japan, sales start today' was of a completely different kind.
SoftBank may proceed with utilizing it in its own mobile phone business (SoftBank/Y!mobile) and expanding it to corporate clients in the future. At the very least, there was enough material in the announcement documents to read it that way.
The wave came faster than I thought. Honestly, that is what I think right now.
Divided into those whose jobs are taken by AI and those who raise AI.
It is easy to read this news and complain, 'Big companies are nice' or 'My call center doesn't have a system budget.' In fact, even I am envious. From the perspective of those of us who are tied up in knots with security and compliance, and who can only manage to secretly try things outside the company by paying 10,000 yen a month out of our own pockets for AI, SoftBank's massive infrastructure and base of over 400,000 corporate clients look like a world far away.
However, what I want to say here is not that we should 'resist anyway'.
Even if you resist, the 97% figure will not change.
Instead, I believe we have no choice but to reshape the role of the SV itself.
The job of an SV until now has been 'managing people.' Creating shifts, checking response quality, taking over difficult calls, and training new staff. The job of an SV from now on must change to 'managing both people and AI agents.'
Sierra co-founder Bret Taylor says that his development philosophy was greatly influenced by the Japanese spirit of 'omotenashi' (hospitality). Who exactly will design and train AI agents to anticipate customer needs and respond with sincerity?
Is it an IT consultant in Tokyo? Is it a brilliant engineer in America?
No. It is us, the people on the front lines, who have faced the angry voices and confused expressions of customers every day and fought against complex business rules. 'In this situation, this is how you speak to show hospitality,' or 'Behind this procedure, there is this business trap.' Incorporating such raw, on-the-ground knowledge into AI, creating rules for judgment, and correcting the course when the AI's answer is off-target—this role is what will become the core business of the SV from now on.
Enthusiasm leaves with the person. Systems remain.
If you try to run a site solely on individual enthusiasm and skills, the site will collapse the moment that person quits. But if we can leave our 'knowledge of hospitality' in the AI as a system, the organization can move to the next stage.
If you are also in a position to manage people on the front lines, this should not be someone else's problem. People who handle common inquiries, people who check procedures, and managers who only know past rules—if they continue to work the way they do now, they will sooner or later be swept away by the same wave. I believe we need to gradually restructure our roles from 'people who manage' to 'people who teach AI the wisdom of the front lines and supervise its judgments' while there is still time.
Starting today, beginning with what we have in our hands
I believe this partnership between SoftBank and Sierra is a symbol that the call center industry has finished its trial period of 'how to use generative AI' and has entered a phase of determination to 'put AI agents into full-scale operation.'
There is no need to wait for large systems or budgets. Instead of giving up by saying 'the company won't give me the tools,' use the free tools at hand or familiar automation to try to systematize your work, starting from a level you can do in under 10 minutes. Invest a little of your own money and time, and burn your bridges to try it out.
Only the accumulation of those gritty steps will become the weapon that carves out the future for me, at 58, and for all of us working on the front lines.
Let's fight on the front lines together.
■ Recommended Reading
If you read this article and felt, 'What exactly is scary about Sierra?' or 'What should I start with on the front lines?', please read the two articles I wrote previously. Both were written inspired by articles by Ryohei Katayama, the founder of paiza.
#GenerativeAI #DXPromotion #CallCenter #Management #OrganizationalDesign #NonEngineer #AIPromotion #Supervisor #SoftBank #Sierra