Spintegrate, Inc. × DBA Japan · Companion-Style AI Support for an Executive in HR and Recruiting · Case Study
Spintegrate, Inc.
HR, Recruiting, and Organizational Consulting · Takuro Ishikura, Executive (Enzo Ishikura)
Takuro Ishikura is an executive who has held recruiting and business leadership roles across Asia Pacific. We provided AI support built around Claude Code, working alongside him rather than lecturing. This case study records what changed in about 1 week after adoption, in his own words and his own estimates.
One week in, management
reporting looked different.
Every number below is a self-reported estimate Mr. Ishikura shared in the first days of adoption.
60 to 70% less time
Time to build management P&L reports
2 prompts
P&L nearly final
Voice input only
No keyboard · zero input errors
Source: Mr. Ishikura's own statements in an interview about one week after adoption (July 2026). These are not third-party measurements.
Voice
“A passion that lights
a fire in people.”
Takuro Ishikura (Enzo Ishikura) / Spintegrate, Inc.
“Kawashima's strength is passion. You rarely meet anyone who moves with that much energy, saying ‘I'll do it. This is what I want to do. What do you say?’”
“He actually founded a company, built things, brought in revenue, and even after all that, he's still hungry. It's not about how quickly he picks things up. He has, how do I put it, a passion that lights a fire in people.”
“I don't think that's something you can learn later in life. To some degree, it's something you're born with.”
Story
What happened
in the first week.
-
Step 1
Moving the management P&L from Excel to Claude Code.
The first project was rebuilding the profit and loss report at the core of his management routine. His reaction to the first output:
“That one really surprised me. It was astonishingly accurate. Clean, exactly the way I asked for it.”
-
Step 2
Two prompts. After that, just updates.
People tend to assume this takes many rounds of back and forth. In his experience, it didn't.
“After about two rounds of instructions, everything came out clean, and from there I was just updating it. If you give it proper instructions, it really lands.”
“It was definitely faster than doing it myself.”
-
Step 3
Voice input only. A way of working without a keyboard.
He likes being able to think while he talks, he says. Not a single input error so far (self-reported).
“I've only used voice input. Honestly, I don't even touch the keyboard anymore. It's just completely easier for me.”
-
Step 4
A first feel for agentic AI.
It doesn't just wait for instructions. It goes and finds the files it needs on its own. That experience stood out to him.
“The way it went off and found the Excel file by itself really surprised me. Living inside the terminal is the best part. It connects with the PC extremely well.”
“So far, nothing has fallen short of my expectations.”
All quotes are Mr. Ishikura's actual statements from an interview and training sessions about one week after adoption.
Not training that lectures.
A companion that moves with you.
DBA Japan's companion-style AI support is not one-way instruction in generative AI tools. We start by interviewing your team to understand your business challenges and current skill level, then design the engagement so people keep using AI on the ground long after the support ends. Learn more on the companion-style AI training page (Japanese).
In your business too,
start with one week of change.
Management reporting, document creation, recruiting operations. Where AI should enter first is different for every company. We will listen to your situation and propose a first step that fits.
Book a Free ConsultationFrequently Asked Questions
What changed through the companion-style AI support at Spintegrate, Inc.?
About one week after adopting Claude Code, the time spent building management P&L reports fell by roughly 60 to 70%, by Mr. Ishikura's own estimate. The P&L was nearly final after 2 prompts, he ran everything on voice input only, and he reported zero input errors so far (all self-reported).
Who measured the “about 60 to 70% less time” figure, and how?
It is not a third-party measurement. In an interview about one week after adoption, Mr. Ishikura said, “I'd say I cut it by around 70%. At the very least, I think it's more than 50 or 60%.” We publish that statement as the approximate figure he gave, as his own early-stage estimate.
How is companion-style AI support different from a typical AI training course?
Instead of one-way instruction in generative AI tools, we start by interviewing your team to understand your business challenges and skill level, then design the engagement so people keep using AI on the ground after the support ends. Learn more on the companion-style AI training page (Japanese) and in the Fujitsu case study.
More Case Studies