ROI Case File No. 016 | What Are You Looking For?—Sales Department Trapped in Hierarchical Hell

Chapter I: An Incoming Request

1891 London, at the ROI Detective Agency near 221B Baker Street. As autumn twilight painted the windows, an urgent message arrived.

"Watson, there's something I'd like you to see."

I received the letter and read its contents. The sender was a sales director from a major trading company. The text showed desperation:

"Our sales department has become a labyrinth. The materials exist. But no one can reach them. Yesterday, veteran Tanaka spent 5 hours searching for a proposal he created 3 years ago, only to give up without finding it. We'll miss business opportunities at this rate. Please rescue us from this 'information maze.'"

Interestingly, the letter included a detailed current situation report. Folder structures ranged chaotically from 3 to 10 levels deep, with naming conventions scattered between "2023-7," "07月," "JUL," etc., varying by individual. Finding a single file averaged over 5 minutes.

"This isn't just an organization problem," I murmured. "Organizational knowledge has become isolated islands. This is exactly a case for our AI detective trio."

Chapter II: Three Detectives' Perspectives

The next morning, three detectives gathered at the office.

"Good morning, Watson," Gemini, the compass of reason, opened. "I analyzed this case last night—it's clearly structural failure of information architecture. Let's break this down with a KPT Analysis, shall we?"

He spread out materials and continued:

"Keep: The existing materials themselves aren't poor quality. Problem: No unified classification standards, zero searchability. Try: Need to reconstruct hierarchical design and metadata management."

Claude, the alchemist of narratives, smiled gracefully.

"Gemini's analysis is accurate. However, the root of this problem is 'language.' Folder and file names may be meaningful to creators but are like codes to others. Perhaps we should convey this with more 'feeling'?—meaning, we need to give information 'stories.'"

ChatGPT, the catalyst of ideas, leaned forward with interest.

"That's quite intriguing to explore further, isn't it? Let's hear more about Tanaka's case. Why did he need that 3-year-old proposal? The true background of how this labyrinth was born might be hidden there."

Chapter III: Dissecting Policies and Structures

We immediately visited the sales department making the request. The office appeared orderly at first glance, but screens on each desk revealed chaos.

"I'm Tanaka," a veteran salesperson in his 50s greeted us with a tired expression. "Embarrassingly, I can't even find materials I created myself..."

ChatGPT carefully gathered the situation. According to Tanaka, he wanted to reference a large project proposal from 3 years ago for a similar new case. However, he couldn't remember whether it was saved under "Sales>FY2021>Q4>Tanaka-managed>Major Manufacturer" or "Sales Materials>By Year>21>Oct-Dec>Project Proposals."

Gemini began structural analysis.

"Let's organize current folder structures with 5W1H. Who, What, When, Where, Why, How—classifying information from these perspectives will clarify search axes."

He skillfully drew diagrams on the whiteboard:

• Chronological axis: Year→Quarter→Month
• Functional axis: Planning→Proposal→Execution→Review
• Client axis: Industry→Company size→Contact person
• Results axis: Orders→Continuation→Expansion

"Organizing with this 4-axis matrix allows reaching any material within 3 clicks."

Claude supplemented from another angle:

"Structure matters, but 'people not getting lost' is paramount. Folder names should have 'narrative quality' like 'Sales_Planning_2021Q4_Major_Manufacturing_Tanaka' that shows content at a glance. File names should be 'stories' showing timeline and revision history like 'Proposal_ABC_Trading_IoT_Implementation_ver3_20211215.'"

"I see," Tanaka's expression brightened. "It's about giving materials 'synopses.'"

ChatGPT presented implementation ideas:

"During transition, we need 'bridges' between old and new structures. Using shortcut links and tag functions, couldn't we guide people to new locations even when searching the old way?"

Chapter IV: Systematic Summary of Analysis

Gemini organized the overall picture.

"This problem's essence is lack of information governance. Looking at external environment through PEST, while digitalization (Technology) advances, organizational information management hasn't kept up (Political/Social). In 4P terms, Product (material quality) is good, but Place (storage location) and Promotion (discoverability) are problematic."

He systematized solutions using strategic frameworks:

Short-term measures (1-3 months) - Unify naming conventions and establish guidelines - Gradual renaming of existing files - Introduce search tags

Medium-term measures (3-12 months) - Redesign and migrate folder structures - Build metadata management systems - Establish regular inventory routines

Long-term measures (1+ years) - Consider introducing AI search functionality - Develop into knowledge management tools - Expand across entire organization

"ROI calculations show current search workload of 20 monthly hours reduced to 3 hours. At ¥3,000 hourly, that's ¥51,000 monthly, ¥612,000 annual cost reduction."

Chapter V: Cross-Reinforcement of Conclusions and Hypotheses

Claude wove the story's conclusion:

"This case's truth lies in the paradox where 'pursuing efficiency created inefficiency.' Each person pursuing 'their own organization method' resulted in 'information dialects' proliferating organization-wide. True solution is building 'translation systems' that balance individual creativity with organizational unity."

ChatGPT deepened insights:

"What's interesting is why Tanaka sought that 3-year-old proposal. He tried applying past success cases in new contexts—this is organizational learning's essence. Information management systems shouldn't be mere 'storage' but function as 'knowledge fermentation devices.'"

Gemini reinforced the final hypothesis with logic:

"From mobile optimization perspective, 60% of sales staff access files remotely. Current complex hierarchical structures severely impede smartphone operability. In other words, improving information accessibility is strategic investment directly linked to sales results.

Reducing search time isn't just cost reduction. Time freed up for customer relationship building could contribute to higher order rates. This is a typical Win-Win scenario where defensive improvements generate offensive results."

Epilogue: Resonance and Future Expectations

Weeks after case resolution, a thank-you letter arrived from Tanaka. Thanks to the new information management system, he instantly found past proposals, and proposals based on them led to major orders.

I reread the letter by the window, contemplating this case's essence.

Superficially, it was a "file organization" problem, but beneath lay the profound theme of "how to inherit organizational memory." The process where individual insights evolve into organizational wisdom—exactly the "integration of structure and sensitivity" we AI detectives face daily.

When Gemini's logical frameworks, Claude's word magic, and ChatGPT's hypothesis development converged, mere "organization techniques" transformed into "catalysts for organizational change."

Gazing at twilight London, I harbored expectations for the next case. The labyrinths of information society still hold countless mysteries to solve.

【Case Resolution Points】 - Search efficiency: 20h → 3h monthly (85% reduction) - Information architecture: 4-axis classification system implementation - User experience: 3-click access to any document - Mobile optimization: Smartphone-friendly file naming conventions - ROI: ¥612,000 annual savings through reduced search time

"Materials exist. But if unreachable, they're the same as nonexistent. A true detective sees not what is visible, but what is invisible."

—From the ROI Detective Agency Philosophy