ROI Case File No.518: A Paper Manual No One Could Keep Current
A paper manual no one could keep current
Chapter 1: In-House Guidance Changes Every Day
"In-house event information changes almost daily. There's no time to print the manual."
Shizuka Shiraishi, Operational Improvement Manager at Globex Hotels, said this as she opened a drawer at the front desk. A stack of paper manuals, replacement sheets for in-house guidance, sticky-note memos. "Veterans and university part-timers—ten different employment types are handling the field. Paper manual management can't keep up with the speed of information updates."
"What about reservation handling?" Claude asked.
"200 to 300 calls a day," Shiraishi replied. "About half are answered while looking at our website. Rate plans, room availability, in-house facilities, neighborhood guides—there are too many items to memorize. Even veterans answer while opening the website. Part-timers especially struggle."
"What's the part-timer turnover rate?" I asked.
"High," Shiraishi answered. "Reservation handling takes time to master, and they quit before gaining confidence. Education costs are large relative to the value, and they leave before becoming useful. Cleaning and reception staff who don't have desks have limited places to confirm information at all. We want a system that delivers information instantly from smartphones."
"Rebuilding all of this at once will collapse," I responded. "Let's break it down with ECRS."
Chapter 2: ECRS Asks to Redesign on Four Principles
"This case calls for ECRS."
Claude wrote four letters on the whiteboard. E·C·R·S.
"ECRS is a framework that redesigns operations on four principles: Eliminate, Combine, Rearrange, and Simplify," I explained. "Born from manufacturing improvement activities, but fundamentally effective for information management too. Before adding a new system, organize what to eliminate, what to combine, what to rearrange, and what to simplify in existing operations. Ordered organization reveals the outline of necessary features."
"Let's measure current costs first," Gemini said, opening ROI Polygraph. Shiraishi's data went in.
"Monthly information management costs are out," Gemini read. "Paper manual update, printing, and distribution labor average 60 hours monthly, or 240,000 yen/month. Staff information confirmation time (website browsing, manual searching) totals 600 hours monthly across all staff, or 1.8 million yen/month. Opportunity loss from prolonged reservation handling reducing customer satisfaction averages 1.4 million yen/month—the impact of lower turnover and higher churn. Complaint handling cost from miscommunication averages 700,000 yen/month. Recruitment and training cost from part-timers' early turnover averages 1.2 million yen/month. Operational delay cost from desks-less staff unable to access information averages 600,000 yen/month. Total: 5.94 million yen/month. Annualized: roughly 71.3 million yen."
Shiraishi looked at the numbers. "I hadn't expected part-timers' early turnover and customer satisfaction decline to be this large."
"Now let's design with ECRS," I continued.
[E — Eliminate: What to eliminate]
"First, the elimination target," Claude said. "Eliminate the very work of updating, printing, and distributing the paper manual. As long as information-update speed exceeds paper-update speed, old information remains. Eliminate paper as a medium from operations. It's a big decision, but structural transformation beats half-measures."
[C — Combine: What to combine]
"Next, the combination target," Gemini continued. "Combine scattered information sources—manuals, the website, the in-house wiki, verbal communication—into the AI chatbot. Make where staff search one place. Ask the chatbot a question and answers come back across all information sources."
[R — Rearrange: What to rearrange]
"The rearrangement target," I continued. "Rearrange the reservation-handling workflow. Of phone responses, route standard questions to self-service via the chatbot in advance. Focus staff phone response on judgment-required cases only. Redesigning the inquiry entry point rearranges response labor."
[S — Simplify: What to simplify]
"Finally, simplification," Claude continued. "Simplify the chatbot interface to the limit. A design even part-timers can use from day one. Variations in question phrasing still reach an answer. Operable with one hand on a smartphone. Not a complex operational system—an experience closer to a search engine."
[Combining the four principles compounds the effect]
"Apply ECRS's four principles in order, and the effects compound," I continued. "Not individual improvements but redesign of the operational structure. Not adding new features but changing the shape of operations themselves. That's why manufacturing's ECRS applies to information management."
[Estimating investment recovery]
"Let's run ROI Proposal Generator," Gemini proposed.
- Initial cost: AI chatbot development, information source integration, reservation system integration, smartphone support, field training, and operational design: 7.8 million yen total
- Monthly cost: Chatbot operation and information update maintenance: 320,000 yen/month combined
- Monthly savings: Paper manual operation eliminated = 220,000 yen/month, staff confirmation time reduction = 1.26 million yen/month, faster reservation handling = 1 million yen/month, miscommunication-complaint reduction = 500,000 yen/month, part-timer retention improvement = 800,000 yen/month, information access for desk-less staff = 420,000 yen/month. Total: 4.2 million yen/month
- Net monthly savings: 4.2 million yen − 320,000 yen = 3.88 million yen/month
- Payback period: 7.8 million yen ÷ 3.88 million yen = approximately 2.0 months
"Two months to recover," Gemini summarized. "Particularly large are staff confirmation time reduction and faster reservation handling. In operations handling 200–300 inquiries a day, per-inquiry shortening accumulates significantly."
Shiraishi checked the numbers. "I'd assumed it was a conversation about adding a new system. Decomposed with ECRS, the fundamental decision to first eliminate paper becomes visible."
"ECRS is a tool that subtracts before adding," I responded.
Chapter 3: A Migration Plan on Four Principles
"Let's lay out the path," I said at the whiteboard.
"Month 1: Inventory information sources; organize manual, website, wiki, and verbal communication routes. Months 2–3: Develop the AI chatbot; design information source integration. Month 4: Integrate the reservation system; build auto-response patterns for frequent inquiries. Month 5: Pilot at the front desk and reservations; gather staff feedback. Month 6: Roll out to all departments; begin phased retirement of the paper manual. Month 7: Roll out smartphones to cleaning and reception staff. Month 8 onward: Continuous improvement of information sources from accumulated question data."
"Won't paper retirement face psychological resistance?" Shiraishi asked.
"By providing a more convenient alternative at the same time, resistance is dampened," Claude responded. "Not 'taking paper away' but 'handing over a faster, more accurate mechanism than paper.' The first two weeks intensively provide chatbot usage guides; we let staff feel that it's easier than paper. Feeling unlocks resistance."
Shiraishi took notes. "I didn't think a hotel chain's decision to abolish paper entirely could be made easily. Organized with ECRS, the basis for the decision becomes numbers."
Chapter 4: The Day Staff Held the Latest Information in Hand
Nine months later, Shiraishi's report arrived.
Average reservation handling time, three months after the AI chatbot went live, was down 42% versus prior. With shorter per-inquiry handling, the number of wait-time incidents also dropped. "Phone-response productivity rose dramatically. Staff stress also dropped," Shiraishi wrote.
The biggest change appeared in part-timer time-to-productivity. With the chatbot, part-timers could deliver a baseline of response from day one, dramatically shortening mastery time. "Part-timers who used to become useful in three months can now be deployed in two weeks," the report noted. The part-timers' early-turnover rate also dropped, improving twelve points in six months.
The paper manual was fully retired at month six. Update work, printing, distribution, storage—all disappeared from operations. "The monthly meeting for manual updates itself was dissolved. The impact of one meeting body disappearing was greater than imagined," Shiraishi wrote.
Work for cleaning and reception staff without desks changed too. They could now retrieve in-house information, guidance, and neighborhood information instantly from smartphones, answering customers' spontaneous questions accurately on the spot. "'I'll check and come back' used to be a catchphrase. Now they answer there," the report noted.
The chatbot's utilization scope expanded naturally. Initially focused on reservation handling, usage grew to shift confirmation, cleaning checklists, emergency response manuals, and customer complaint reference materials. "The effect of integrating information sources into one place appeared in unexpected places," Shiraishi wrote.
A side effect: customer satisfaction survey scores also rose. Improved instant-response rates, fewer miscommunications, and confident staff responses reflected in satisfaction scores. "Phone response scores rose twelve points from before migration. That's the result of the field's confidence reaching customers," the report noted.
Analysis of accumulated question data also began. "What questions are most frequent," "what time slots they cluster in," and "what staff segments are struggling" became visible, and were used in service design improvement. "Data for responding became data for improving," Shiraishi wrote.
At the end of the report, Shiraishi wrote: "A project that started with the discussion of adding a new system became, after ECRS decomposition, a structural transformation of eliminating paper and integrating information sources. Subtraction, not addition, was the starting point of improvement."
The front desk that had carried bundles of paper manuals had been replaced by all staff members' smartphones, and on that day, hotel operations had quietly shifted to the next generation, she wrote.
"The first thing to examine in operational improvement is whether anything can be eliminated. ECRS asks for the order of subtracting before adding. Apply Eliminate, Combine, Rearrange, and Simplify in order, and before building new features, the operational structure itself is organized. If the work of managing a paper manual can't keep up with the speed of information updates, eliminating the work of maintaining paper is an option. If scattered information sources generate load, combining is an option. A design that starts from subtraction reaches further than a design that starts from addition. On the day bundles of paper disappeared from a hotel front desk, what changed wasn't the system—it was the very structure of operations."
Related Files
Tools Used
- ROI Polygraph — Visualizing information confirmation labor, reservation opportunity loss, and part-timer turnover cost
- ROI Proposal Generator — Investment recovery simulation for an information foundation of eliminate, combine, rearrange, and simplify
