ROI Case File No.313 | 'AquaCall's Inaudible Success'

Chapter 1: The Illusion of AI—Implementation Without Results

The week after KAYOU's AARRR case was resolved, a consultation arrived from Tokyo regarding AI utilization at a sales support company. Case File 313 of Volume 26, "The Pursuit of Reproducibility," tells the story of a company where AI implementation didn't yield results, but daily reflection habits made improvement routine.

"Detective, we implemented an AI telemarketing system three months ago. The AI analyzes prospect lists and makes calls at optimal times. However, the meeting conversion rate hasn't improved at all. Of our 8 salespeople, only 2 are producing results. The remaining 6 say 'AI doesn't work.'"

Takashi Yamaguchi, Sales Manager of AquaCall Co., born in Minato Ward, visited 221B Baker Street unable to hide his confusion. In his hands were the AI system implementation plan and, in stark contrast, sales meeting minutes marked "no results."

"We sell sales support tools to small and medium-sized enterprises in Tokyo. Our main product is an integrated CRM and SFA system. Our sales team has 8 members. We conduct about 2,400 telemarketing calls monthly, but the meeting conversion rate is stagnating."

AquaCall's AI Implementation Stagnation: - Established: 2018 (sales support tool sales) - Annual Revenue: 480 million yen - Sales Team: 8 people - Monthly Call Volume: About 2,400 calls (300 per person) - Meeting Conversion Rate: 8% before AI implementation → 8% after (no change) - AI Investment: 4.8 million yen system implementation + 120,000 yen monthly operation fee - Problem: Polarization between salespeople who can use AI and those who cannot

Yamaguchi's voice carried deep anxiety.

"The problem is not knowing how to use AI. The AI says 'Call this customer within this week.' But we don't know why it made that judgment. Salespeople say 'I did what AI said but got rejected.' Because there are no results, an atmosphere of 'AI doesn't work' is spreading."

Typical Salesperson's Day:

Salesperson A (28 years old, 3 years with company):

9:00 AM: Logs into AI system. Today's recommended call list displays.

【AI's Recommended List】
1. Company B (Priority: High)
   Reason: Website visited 3 times, document download history
2. Company C (Priority: Medium)
   Reason: Email opened, link clicked
3. Company D (Priority: Low)
   Reason: 2 months since business card exchange
... (30 total)

9:15 AM: Calls Company B.

"Thank you for your time, this is Salesperson A from AquaCall. I noticed you recently viewed our website..."

Company B Contact: "Ah, I saw it, but we don't need it right now."

Call duration: 45 seconds, Result: Rejected

Salesperson A was puzzled.

"AI said 'Priority: High,' so why was I rejected...?"

9:30 AM - 11:30 AM: Calls remaining 29. Meeting conversions: 1 (3.3%)

After lunch, afternoon: Another recommended list displays. Calls the same way. Gets rejected the same way.

Evening: Today's results: 60 calls, 2 meeting conversions (3.3%)

"AI says 'this is good,' but I don't know why it's good. When rejected, I don't know what was wrong. How should I improve...?"

Meanwhile, Salesperson B (32 years old, 5 years with company) producing results:

Using the same AI system, but with an 18% meeting conversion rate.

Salesperson B's Day:

9:00 AM: Looks at AI's recommended list but doesn't call immediately.

"First, I think about why AI recommended this customer. If there's website visit history, I check which pages were viewed. If there's document download, I check which document."

9:15 AM: Before calling Company B, preparation.

• Checks Company B's website: "Operational efficiency" theme
• Downloaded document: "Sales management automation"
• Hypothesis: "Perhaps they have issues with sales management?"

Call: "Thank you for your time, this is Salesperson B from AquaCall. You recently downloaded our 'Sales Management Automation' document. Are you currently experiencing any challenges with sales management?"

Company B Contact: "Ah, actually our sales case management is person-dependent..."

Meeting conversion successful

Evening: Today's results: 60 calls, 11 meeting conversions (18%)

Salesperson B takes daily notes on call content.

"Success pattern: For customers who downloaded documents, asking questions about that document's theme expands the conversation."

Chapter 2: The Wall of Tacit Knowledge—Successful Methods Aren't Shared

"Yamaguchi-san, what do you think causes the difference in meeting conversion rates between Salesperson B and Salesperson A?"

To my question, Yamaguchi answered.

"B has extensive experience. He has the ability to read customer reactions. On the other hand, A is still young. He lacks experience. However, when I ask B 'How do you do it?', he only says 'Experience.' How can I teach that experience to A...?"

Current Approach (Person-Dependent Type): - Results: Depend on individual experience and intuition - Sharing: Only verbal explanations of "I do it this way" - Problem: Success patterns aren't verbalized or shared

I explained the importance of making improvement a habit.

"Don't lock up experience. KPT—Keep, Problem, Try. What to continue, what's the problem, what to try. By repeating these three, success patterns become organizational assets."

⬜️ ChatGPT | Catalyst of Conception

"Don't lock experience within one person. Verbalize success with KPT and transform it into everyone's weapon."

🟧 Claude | Alchemist of Narrative

"Improvement doesn't dwell in meeting rooms but in daily life. KPT transforms reflection into habit."

🟦 Gemini | Compass of Reason

"KPT is the mirror of improvement. Reproduce results through the 3 stages of Keep (continue), Problem (issue), and Try (challenge)."

The three members began their analysis. Gemini unfolded the "KPT Framework" on the whiteboard.

KPT's 3 Steps: 1. Keep: Continue what worked well 2. Problem: Identify what didn't work well 3. Try: Decide what to try next

"Yamaguchi-san, let's introduce KPT habits to AquaCall's sales team."

Chapter 3: The Daily Mirror—15 Minutes of Reflection Creates Results

Phase 1: KPT Introduction (1 week)

First, we explained the KPT mechanism to all sales team members.

Rules: - Every day, write on KPT sheet for 15 minutes before end of work - Next morning, share with entire team for 5 minutes - Weekly, 30-minute KPT reflection meeting

KPT Sheet (Template):

【Date】November 12, 2025
【Person】Salesperson A

【Keep (what went well)】
- 

【Problem (what didn't go well)】
- 

【Try (what to try next)】
- 

Initially, the sales team was bewildered.

"What should we write?"

"What went well, what didn't go well, what to try next. That's all. For example..."

Phase 2: Week 1 KPT Practice

Salesperson A's Day 1 KPT:

【Keep】
- Focused on AI's top 5 recommendations
- Finished calls in the morning

【Problem】
- Didn't understand AI's recommendation reasons, called without preparation
- 2 meeting conversions out of 60 calls (3.3%)

【Try】
- Tomorrow, I'll check AI's recommendation reasons before calling

Salesperson B's Day 1 KPT:

【Keep】
- For document download customers, asked questions about that document's theme (high success rate)
- Inferred customer interest from website visit history

【Problem】
- Customers with only email opens don't expand conversation (get rejected)

【Try】
- For email open customers, first ask "which part interested you"

Next Morning's 5-Minute Sharing:

Salesperson A was surprised to see B's sheet.

"B-san, 'ask questions about theme for document download customers'—that's what it meant! I was calling without thinking..."

Salesperson B: "That's right. If you look at AI's recommendation reasons, you understand customer interest. If you talk matching that, meeting conversion is easier."

Salesperson A: "I'll try it from today."

Phase 3: Week 2 KPT Practice and Improvement

Salesperson A's Day 7 KPT:

【Keep】
- Check AI's recommendation reasons before calling (meeting conversion rate 3.3% → 8%)
- For document download customers, ask questions about that theme

【Problem】
- Customers with only website visits still often reject

【Try】
- Check "which pages" they viewed on website (B-san's method)

Salesperson B's Day 7 KPT:

【Keep】
- Asking email open customers "what part interested you" (success rate improved)

【Problem】
- Many customers don't answer phone (call timing issue?)

【Try】
- Instead of AI's recommended time slot, try optimal time by industry
  (Manufacturing: 10 AM, IT: 3 PM)

All sales team members write KPT daily and share the next morning.

Then, success patterns gradually get verbalized and shared across the team.

Phase 4: KPT Analysis After 1 Month

We aggregated 1 month of KPT sheets and extracted success patterns.

Team-Wide Keep (Success Patterns) Top 5:

1. For document download customers, ask questions about that theme (Mentions: 32 times)
2. Meeting conversion rate: 18%
3. Check website visit pages and infer interest (28 times)
4. Meeting conversion rate: 15%
5. Ask email open customers 'what part interested you' (24 times)
6. Meeting conversion rate: 12%
7. Call at optimal time by industry (18 times)
8. Phone connection rate: 38% → 52%
9. Check AI's recommendation reasons in advance (48 times)
10. Meeting conversion rate: 8% → 12%

Documented these as "Sales Knowledge Collection" and distributed to everyone

Team-Wide Problem (Common Issues) Top 3:

1. Customers 2+ months after business card exchange don't expand conversation (Mentions: 42 times)
2. Customers with only email opens are easily rejected (38 times)
3. Can't reach decision-makers (28 times)

Discussed Try (initiatives to test) for these issues with everyone

Phase 5: AI and KPT Integration (Month 2)

We fed success patterns accumulated through KPT back into the AI system.

Requested from AI development vendor: - Add the 5 success patterns to AI's recommendation logic - When AI presents recommendation reasons, clearly indicate "document download customer," "website visit pages," etc.

Month 2's New AI Recommendation Screen:

【AI's Recommended List】
1. Company B (Priority: High)
   Reason: Document download (sales management automation)
   → Recommended talk: "Are you experiencing any challenges with sales management?"

2. Company C (Priority: Medium)
   Reason: Website visit (pricing page viewed)
   → Recommended talk: "Do you have questions about implementation costs?"

Salesperson A's reaction: "AI tells me 'why' and 'how to talk'! Even I can do this!"

Chapter 4: Habit as Asset—Organizational Change After 6 Months

Results After 6 Months:

Dramatic Improvement in Meeting Conversion Rate: - Entire sales team: 8% → 16% (+100%) - Salesperson A (was 3.3%): 3.3% → 14% - Salesperson B (was 18%): 18% → 22% - Other 6 people also converged to average 15%

Financial Impact: - Monthly meeting conversions: 192 cases → 384 cases (+100%) - Conversion rate (meeting → order): 25% (unchanged) - Monthly orders: 48 cases → 96 cases - Average order value: 1.2 million yen - Monthly sales: 57.6 million yen → 115.2 million yen (+100%) - AI investment recovery period: 2.1 months

Organizational Change:

KPT Habit Formation: - Daily 15-minute KPT writing established (absence rate: 2%) - Morning 5-minute sharing became culture - Weekly reflection meetings update monthly success patterns

Knowledge Accumulation: - 720 cumulative KPT sheets in 6 months - Success patterns: 38 documented - New salesperson ramp-up period: 6 months → 2 months

Yamaguchi's Reflection:

"When we first implemented AI, I thought 'AI will solve everything.' But AI is just a tool. If the people using it don't grow, there are no results.

By introducing KPT, the sales team began verbalizing 'why it succeeded' and 'why it failed.' And we became able to share success patterns with everyone.

Now, even new hires become productive in 2 months. Veteran experience has become an organizational asset."

Chapter 5: The Detective's Diagnosis—Improvement Dwells Not in Meeting Rooms But in Daily Life

That night, I contemplated the essence of KPT.

AquaCall implemented AI but didn't get results. AI presented recommended lists but didn't teach "why" it recommended or "how" to talk.

By making daily reflection a habit with KPT, success patterns were verbalized. And they were shared with everyone and became organizational assets.

"Improvement isn't born in a single meeting. It's born by reflecting small and trying small in daily life. That accumulation changes the organization."

The next case will also depict the moment when KPT makes improvement routine.

"Don't lock up experience. Reflect daily with KPT. Keep, Problem, Try. These 3 steps transform person-dependency into organizational knowledge"—From the Detective's Notes

Related Files