ROI Case File No.473 'Why a Five-Hundred-Person Organization Ground to a Halt'
Why a Five-Hundred-Person Organization Ground to a Halt
Chapter 1: Three People Reading Newspapers All Day, Every Day
"Three staff members are devoted to newspaper clipping—all day, every day."
Kenji Fujimoto, Head of Corporate Planning at TechNova, spread a set of documents on the table as he spoke. This year's AI utilization promotion plan, and a list of operational challenges submitted by each department. Four sheets of paper, layered.
"What does the clipping work actually involve?" I asked.
"Every morning, they check multiple newspapers and web media, pull out articles related to our company, competitors, and industry, compile them, and send them to the relevant executives," Fujimoto answered. "Some print newspapers get scanned. Web articles get copied and pasted into Word. Because the format isn't unified, there's always a final adjustment step. Three people spend from morning to early afternoon doing nothing but that."
"There are other challenges on the list as well," Claude said, scanning the document.
"There are," Fujimoto continued. "Customer inquiries only come in through phone and email—if the person in charge is out, we call back. App inquiries are often the same questions repeated; they'd resolve themselves with a proper FAQ, but every time we put a human on it. E-commerce inquiries are 80% shipping-related, the answers are fixed, and yet we type individual replies every time."
"What about customer data analysis?" Gemini asked.
"Not happening," Fujimoto answered immediately. "The data lives in the core system, but no one can analyze it. Demand forecasting relies on experienced staff's intuition."
"So there are five challenges in total," I summarized. "Clipping, customer service, app FAQ, e-commerce inquiries, and data analysis. Each has been accumulating in isolation."
Fujimoto nodded. "We can't prioritize. Management is pushing for faster AI adoption, but the field doesn't know how to move."
"We need to look at the structure of the whole organization," I said.
Chapter 2: The Seven Alignments of 7S
"This case calls for 7S."
Claude wrote seven words on the whiteboard: Strategy, Structure, Systems, Shared Values, Skills, Style, Staff.
"7S is an organizational analysis framework developed by McKinsey," I explained. "You can't solve organizational problems by changing just one element. Whether a transformation succeeds or fails depends on how well all seven elements interlock. Rushing AI adoption by changing only Systems while the other six don't keep up—that's the structure that produces systems that go unused."
"Let's start by measuring the current cost," Gemini said, opening ROI Polygraph. Work logs and department-level time data from Fujimoto were entered.
"The monthly cost of the five challenges is in," Gemini read aloud. "Clipping: three staff × 22 days × 6 hours × ¥2,600/hr = ¥1,027,200/month. Customer service callback and wait costs: est. ¥300,000/month. App FAQ handling: ¥150,000/month. E-commerce manual replies: ¥220,000/month. Opportunity cost from absent data analysis: deferred due to difficulty of estimation. Combined total: ¥1,697,200/month. Annualized: ¥20,366,400."
Fujimoto looked up from the documents. "So clipping alone is over ¥12 million a year on a headcount basis?"
"That's the calculation," Gemini confirmed.
"Now let's work through 7S," I continued.
[Strategy — Define AI Utilization in a Single Sentence]
"TechNova currently has no verbalized AI strategy," Claude said. "Being told by a parent company to accelerate adoption is an instruction, not a strategy. Strategy is the answer to why you're using AI."
"We may never have asked ourselves why," Fujimoto said.
"Let's write it in one sentence today," I proposed. "TechNova uses AI in order to maintain operational quality without additional headcount—and to increase the speed of customer response and decision-making. That sentence is the strategy. All five challenges connect to it."
[Structure — Who Is Responsible for AI Adoption?]
"Who currently owns AI adoption?" Claude asked.
"Loosely speaking, I do—but not exclusively," Fujimoto answered. "I carry my corporate planning responsibilities while trying to move cross-functionally."
"That's the Structure problem," Gemini said. "When the responsible person isn't dedicated, priority conflicts arise every day—core work versus AI advancement. For the organization to move forward, securing even one dedicated resource for the AI project comes first."
[Systems — Which System Do We Touch First?]
"Of the five challenges, the first one to address is clipping," I said. "Three reasons: it has the largest impact, the lowest technical difficulty, and the most visible outcome. If clipping automation succeeds, the organization develops a felt sense of what AI can do. That felt sense creates buy-in for the next initiative."
"Today is the first time I've had a basis for sequencing," Fujimoto said.
[Shared Values — How Do We Handle Resistance to AI?]
"Is there resistance to AI in the field?" I asked.
"Yes," Fujimoto answered immediately. "Particularly among longer-tenured staff—there's anxiety that jobs will be taken."
"Shared Values design is needed," Claude said. "What AI replaces is repetitive tasks. The three clipping staff will move into roles that evaluate the accuracy of articles AI surfaces and consider how to frame them for executives. It's not about being replaced—it's about moving into higher-value work. Communicating that in writing before implementation matters."
[Skills, Style, Staff — Preparing the People]
"The remaining three elements all relate to people," Gemini summarized. "For Skills: one round of foundational AI training for all staff. For Style: explicitly build a culture where humans review and approve AI-generated outputs. For Staff: identify the one most AI-enthusiastic person in the organization and make them the pilot user—let their success story spread naturally."
"Let's run an investment simulation through ROI Proposal Generator," Gemini proposed.
A phased implementation projection anchored by clipping automation was produced.
- Initial cost: Clipping automation + FAQ build + e-commerce auto-reply setup — ¥1,500,000
- Monthly cost: Combined system subscriptions — ¥150,000/month
- Monthly savings: 80% reduction in clipping (3 staff) = ¥820,000; FAQ + e-commerce automation = ¥370,000; total = ¥1,190,000/month
- Net monthly savings: ¥1,190,000 − ¥150,000 = ¥1,040,000/month
- Payback period: ¥1,500,000 ÷ ¥1,040,000 = approx. 1.4 months
"Payback in one and a half months," Gemini summarized. "Moving with 7S to prepare the organization prevents the failure mode of systems racing ahead of people."
Chapter 3: The Day the Sequence Was Decided
"Let me lay out the plan," I said, standing at the whiteboard.
"Month one—verbalize the strategy and secure a dedicated resource. Reach agreement with leadership on the one-sentence AI strategy and assign one dedicated person. Month two—pilot clipping automation. Test against one week of articles first to verify precision, then define the new roles for the three staff. Month three—deploy FAQ and e-commerce auto-reply. Share the clipping success story internally before starting. Month four onward—build the customer data analysis foundation. The reason this comes last: it has the highest technical difficulty, and without prior success experience to draw on, the risk of stalling is high."
"If we follow this sequence, will we avoid failure?" Fujimoto asked.
"There's no sequence that avoids failure," I answered. "But there is a sequence that lets you learn from failure. Try small first, expand when it works. 7S is the design that makes that expansion mesh across the whole organization."
Fujimoto gathered his documents. "I'd been trying to move all five at once. Today, the sequence got decided."
Chapter 4: The Day the Three Attended the Executive Meeting
Nine months later, a report arrived from Fujimoto.
Clipping automation stabilized within three weeks of going live. The system now handles article retrieval, and the three staff members moved into roles handling content verification and editorial summary. Work that had consumed the entire morning was completing before noon.
One of the three, using her freed-up time, began producing a competitive analysis report. "An executive asked me to keep doing it every week," Fujimoto wrote in his report.
FAQ and e-commerce auto-reply launched in month three. App inquiry volume dropped 43% month-over-month. 81% of e-commerce shipping inquiries were handled automatically.
Internal resistance to AI softened quickly, Fujimoto noted, because the visible transformation of the three clipping staff became a concrete case study the organization could point to.
Seven elements had started, slowly, to click into place.
"Organizations usually can't use AI because of technology—they fail because the strategy isn't verbalized, the responsible person isn't dedicated, the field has resistance, and the skills aren't there. If any one of the seven elements is missing, systems race ahead alone. What 7S asks isn't what to install—it's what isn't aligned. On the day the three clipping staff attended an executive meeting, five hundred people got a little lighter."
Related Files
Tools Used
- ROI Polygraph — Visualizing monthly costs across five challenges
- ROI Proposal Generator — Simulating ROI on phased AI implementation
