AgenticForce Logo
AgenticForce
🧠 Case Study

Smiley
Automated Customer Service

How we reduced response time from 4 days to 15 minutes with AI

-99%
Response time
94.5%
CSAT Score
+15%
Sales conversion
🎯 Challenge

Problems to solve

Smiley is a dynamically growing e-commerce store handling hundreds of inquiries weekly

Long response time

4 days

Average response time to customer inquiries

Inconsistent quality

Low

Response quality dependent on agent

No central system

0%

Service process automation

Project goals

Reduce response time
AI-based support system
Response standardization
24/7 full workflow
🛠️ Solution

System architecture

Complete, agent-based customer service system running 24/7

Form

Dedicated frontend

n8n

Automation engine

Data

Coda + Baselinker

GPT-4

Response generation

Contact form

Frontend with tagging

Dedicated frontend with topic tagging, integration with context database

n8n

Automation engine

Handles incoming inquiries, recognizes topics, retrieves data and launches AI

Coda

Baza wiedzy

Baza wiedzy i stylu komunikacji (tone of voice, gotowe elementy odpowiedzi, polityki)

GPT-4

AI Engine

Generuje odpowiedzi dopasowane do klienta, stylu i historii

Baselinker

E-commerce Hub

Entry point + message sending + order status synchronization

Ticket System

Case Tracking

Creates, tracks and assigns cases with unique ID and status

⚙️ Workflow

How it works

Complete flow from request to resolution in 6 steps

1

Customer inquiry

Customer sends inquiry through dedicated form or email → automatic routing to n8n

Form Email Auto-routing
2

Automatic processing

n8n recognizes the topic (lost package, return, complaint), fetches data from Baselinker and history from Coda

Classification Data fetching Context
3

Ticket creation

System creates unique ticket with ID and status for efficiency monitoring and SLA

Unique ID Status tracking SLA monitoring
4

AI generates response

GPT-4 creates personalized response aligned with company policy and customer context

GPT-4 Personalization Brand voice
5

Response delivery

Message sent automatically via email or as draft for agent approval

Auto-send Manual review Email
6

Follow-up and closure

Ticket update, follow-up assignment or automatic case closure

Update status Follow-up Auto-close
📊 Results

Business impact

Spectacular results in every key metric

Metric Before implementation After implementation Change
Average response time 4 days 15 minutes −99%
Customer satisfaction (CSAT) 58% 94.5% +63%
Manual agent work 100% 40% −60%
Support sales conversion baseline +15% +15%
Interactions per case 2–4 1–2 ✅ Less frustration
Response consistency Low, agent-dependent High, brand-aligned
-99%

Response time

From 4 days to 15 minutes

94.5%

CSAT Score

63 point increase

-60%

Manual work

Agents focus on complex cases

+15%

Conversion

More sales from support

🚀 Future

Development directions

Next steps in customer service automation

Channel expansion

Additional communication channels for full omnichannel experience

Messenger WhatsApp Live Chat

AI learning

Continuous improvement based on tags and feedback

Machine Learning Feedback Loop Auto-improve

KPI Dashboard

Real-time monitoring of all customer service metrics

Real-time CSAT Response Time

SLA Alerts

Automatic escalation when time limits are exceeded

20min SLA Auto-escalation Alerts

Want similar results?

Let's discuss customer service automation for your company. Free consultation and process analysis.