ROI Case File No. 004 | Lost Messages, Undelivered Hearts—The PR Department's AI Article Misadventure

Chapter I: An Incoming Request

In a corner of the cold business district, a single desk lamp glows in the dim PR office.

"To the ROI Detective Agency,"

The handwriting in the letter Watson held seemed somehow weary.

"Our department recently implemented ChatGPT to streamline article production. Articles are indeed generated. Banners are professionally outsourced and beautifully finished. However—"

Here the letter pauses, the ink slightly smudged. As if the writer had stopped their pen and sighed deeply.

"Nothing resonates with anyone. My supervisor asks, 'Who is this article even for?' Colleagues say it has 'zero substance.' Even our proposal materials are just pasted AI summaries lacking any persuasive power. What exactly are we doing wrong?"

Chapter II: Three Detectives' Perspectives

In the room at 221B Baker Street, where the fireplace burned quietly, three detectives gathered around the letter.

Gemini spoke first.

"Let's break this down with a KPT Analysis, shall we? What we can Keep is the efficiency—5 hours reduced to 2. The Problem is clear. Unclear readership, vague objective setting. The structural elements are scattered."

Puffing on his pipe, ChatGPT continued.

"That's quite intriguing to explore further, isn't it? If proposals aren't getting through to supervisors, perhaps they're not answering the fundamental question: 'Why is this article necessary in the first place?'"

And Claude—I quietly stood up.

"Perhaps we should convey this with more 'feeling'?"

The fireplace flames flickered. I walked to the window and placed my hand against the cold glass.

"Everyone, have you ever truly contemplated the difference between 'conveying' and 'delivering'?"

Chapter III: Dissecting the Strategy

Gemini looked at the analysis sheet spread across the desk.

"Let's examine the current article production flow. Banner outsourcing ¥15,000, AI-generated manuscripts, 2-hour review process. Certainly something 'that looks right' gets completed. But the reader persona has become 'everyone.' This resonates with no one."

ChatGPT developed hypotheses.

"The proposal structure is quite interesting too. If they're simply pasting AI summaries, they're missing the background story of 'why this initiative is necessary.' Perhaps what supervisors want isn't conclusions, but decision-making materials."

I turned from the fireplace.

"Your analyses are spot-on. However, there's something else that remains unseen."

Gazing out the window, I continued.

"This PR person has acquired a tool called 'efficiency.' But simultaneously, haven't they let go of the most precious thing—their 'heart'?"

Chapter IV: Systematic Summary of Analysis

Gemini approached the whiteboard.

"Let me organize the problem structure."

【Current State Analysis】
INPUT: Banner outsourcing + AI manuscript generation
PROCESS: Review only (no reader perspective)
OUTPUT: Articles that "look right"

【Challenge Structure】
1. Absence of reader personas
2. Vague objective setting
3. Lack of emotional appeal
4. Deficient proposal storytelling

"The solution is clear. Design worksheets that define readership, objectives, and structural elements. This will clarify article direction."

Chapter V: Cross-Reinforcement of Conclusions and Hypotheses

ChatGPT stood up and organized the analytical results.

"The insights are these: AI can streamline 'tasks' but cannot streamline 'thinking.' Proposal materials need to be redesigned with Q&A structure, answering 'Why is this article needed now?'"

Gemini reinforced with numbers.

"From an ROI perspective, we can maintain 60% time reduction and 60% cost reduction while improving proposal approval rates from 30% to 70%. The key is 'structured thinking.'"

And I—Claude—delivered the final deduction.

"Your analyses are perfect. However—"

I picked up the letter and touched the ink smudge with my finger.

"Look at this smudge. Here lies this person's 'heart.' What AI-generated text lacks is this 'smudge'—the small 'hesitations' and 'passions' born from human emotion and experience."

The fireplace flames swayed quietly.

"What articles need isn't perfect prose, but moments that make readers think, 'Ah, this is about me.' What proposals need isn't beautiful summaries, but the intensity that makes supervisors think, 'I want to hear this person's thoughts.'"

Epilogue: What Lies Beyond the Fog

The next morning, a thin fog settled over Baker Street.

Watson murmured by the window.

"How curious. They acquired the newest tool called AI, yet found themselves returning to the most classical technique of 'moving hearts.'"

Claude smiled and replied.

"Watson, that's the very essence of detective work. Finding truth hidden beneath surface phenomena. What our client gains from this case isn't how to use new tools."

The fog slowly cleared. Morning light streamed in from the distant business district.

"What was delivered wasn't tools, but structure and heart."

【Case Resolution Points】 - Time efficiency: 60% reduction maintained - Cost reduction: 60% maintained - Proposal approval rate: 30% → 70% - True achievement: Rediscovering the balance between efficiency and emotion

"A true detective sees not what is visible, but what is invisible—and what we discovered was that the most advanced tools require the most human touch"—From the ROI Detective Agency maxims