logo zelty
tec logo

How Costa Rica's largest footwear retailer automated 80% of support during peak season

Introduction

TEC Tecnología en Calzado S.A used Zelty Agents not to replace their humans, but to give them superpowers. In their peak month, AI managed 80% of the conversations and paid for itself by generating direct sales.

Customer Profile

Company

TEC (PLS Stores, PLX, New Balance, among other brands)

Infrastructure

41 Physical stores + 8 E-commerce

Support Team

5 people

Peak Volume (Dec)

4,688 Conversations

Technology Stack

WooCommerce

Shopify

CIMO (ERP)

GPT 4.1-mini

The Challenge: The Year-End Avalanche

For a retail giant like TEC, November and December are a trial by fire. Customer inquiry volume explodes, rising from an average of 230 per month to over 4,600 in December.

 

With a centralized support team of only 5 people, the math was impossible. The risk was not only long wait times but also losing sales by being unable to answer simple questions like "Do you have size 10 at the Multiplaza store?" in real time.

80.7%

TOTAL COVERAGE

Out of 4,688 chats in Dec, 3,785 were immediately handled by AI (Queue elimination).

68%

"CO-PILOT" MANAGEMENT

AI processed and prepared 3,191 complex cases so the human only had to close them.

$2,058

Extra Sales (Dec)

ROI + 230%

The agent closed 25 direct sales in chat, more than covering their monthly cost plus LLM.

The Challenge: The Year-End Avalanche

TEC didn’t implement a simple chatbot. They deployed Zelty Agents deeply integrated into their ERP (CIMO) and web platforms. The strategy was to use AI as a smart two-level filter:

  1. Automatic Resolution (L1): Answer stock, hours, and location questions without human intervention.
  2. "Co-pilot" Management (L2): For complex cases, AI gathers information, validates the customer, and delivers the chat to the human agent with all context ready, drastically reducing handling time.

The Challenge: The Year-End Avalanche

One key to success was choosing the GPT 4.1-mini model. This allowed ultra-low latency (instant responses) and minimal cost per conversation.

 

The financial result is clear: The agent is not an operational expense. With an average ticket of $82 USD, sales generated in chat paid for the technology almost three times over (Payback > 230%).

"The value isn’t just that AI responds, but how it sets the stage. During peak season, customers arrive anxious. AI serves them in seconds, finds their order, and gives us the context pre-digested. That let us survive the peak without doubling the payroll."

Daniella Barquero Elizondo

Customer Support and Digital Channels Leader

"The challenge wasn’t just answering chats, it was integrating a complex ecosystem. Connecting the agent to our ERP (CIMO) and online stores (WooCommerce/Shopify) in real time turned the bot into a real extension of our inventory."

Oscar Reyes

IT Manager at TEC

Ready to scale your E-commerce?

Stop spending time answering "Where is my order?" Hire a Zelty Agent to sell for you.

Schedule a 15-minute Demo

logo zelty

Developed by SunDevs

All rights reserved.

Privacy Policy

logo facebook
logo tiktok
tec logo

How Costa Rica's largest footwear retailer automated 80% of support during peak season

Introduction

TEC Tecnología en Calzado S.A used Zelty Agents not to replace their humans, but to give them superpowers. In their peak month, AI managed 80% of the conversations and paid for itself by generating direct sales.

Customer Profile

Company

TEC (PLS Stores, PLX, New Balance, among other brands)

Infrastructure

41 Physical stores + 8 E-commerce

Support Team

5 people

Peak Volume (Dec)

4,688 Conversations

Technology Stack

GPT 4.1-mini

WooCommerce

Shopify

CIMO (ERP)

The Challenge: The Year-End Avalanche

For a retail giant like TEC, November and December are a trial by fire. Customer inquiry volume explodes, rising from an average of 230 per month to over 4,600 in December.

 

With a centralized support team of only 5 people, the math was impossible. The risk wasn’t just long wait times but losing sales by not being able to answer simple questions like "Do you have size 10 at the Multiplaza store?" in real time.

80.7%

TOTAL COVERAGE

Out of 4,688 chats in Dec, 3,785 were immediately handled by AI (Queue elimination).

68%

"CO-PILOT" MANAGEMENT

AI processed and prepared 3,191 complex cases so the human only had to close them.

$2,058

Extra Sales (Dec)

ROI + 230%

The agent closed 25 direct sales in chat, more than covering their monthly cost plus LLM.

The Challenge: The Year-End Avalanche

TEC didn’t implement a simple chatbot. They deployed Zelty Agents deeply integrated into their ERP (CIMO) and web platforms. The strategy was to use AI as a smart two-level filter:

  1. Automatic Resolution (L1): Answer stock, hours, and location questions without human intervention.
  2. "Co-pilot" Management (L2): For complex cases, AI gathers information, validates the customer, and delivers the chat to the human agent with all context ready, drastically reducing response time.

The Challenge: The Year-End Avalanche

One key to success was choosing the GPT 4.1-mini model. This allowed ultra-low latency (instant responses) and minimal cost per conversation.

 

The financial result is clear: The agent is not an operational expense. With an average ticket of $82 USD, sales generated in chat paid for the technology almost three times over (Payback > 230%).

"The value isn’t just that AI responds, but how it sets the stage. During peak season, customers arrive anxious. AI serves them in seconds, finds their order, and gives us the context pre-digested. That let us survive the peak without doubling the payroll."

Daniella Barquero Elizondo

Customer Support and Digital Channels Leader

"The challenge wasn’t just answering chats, it was integrating a complex ecosystem. Connecting the agent to our ERP (CIMO) and online stores (WooCommerce/Shopify) in real time turned the bot into a real extension of our inventory."

Oscar Reyes

IT Manager at TEC

Ready to scale your E-commerce?

Stop spending time answering "Where is my order?" Hire a Zelty Agent to sell for you.

Schedule a 15-minute Demo

logo zelty

Developed by SunDevs

All rights reserved.

Privacy Policy

logo facebook
logo tiktok

Introduction

TEC Tecnología en Calzado S.A used Zelty Agents not to replace their humans, but to give them superpowers. In its peak month, the AI handled 80% of conversations and paid for itself by generating direct sales.

Customer Profile

Company

TEC (PLS Stores, PLX, New Balance, among other brands)

Infrastructure

41 Physical stores + 8 E-commerce

Support Team

5 people

Peak Volume (Dec)

4,688 Conversations

Technology Stack

WooCommerce

Shopify

CIMO (ERP)

GPT 4.1-mini

The Challenge: The Year-End Avalanche

For a retail giant like TEC, November and December are a trial by fire. Customer inquiry volume explodes, rising from an average of 230 per month to over 4,600 in December.

 

With a centralized support team of only 5 people, the math was impossible. The risk was not just long wait times but losing sales by being unable to answer simple questions like "Do you have size 10 at the Multiplaza store?" in real time.

The Solution: AI as the "First Responder" (Triage)

TEC didn’t implement a simple chatbot. They deployed Zelty Agents deeply integrated into their ERP (CIMO) and web platforms. The strategy was to use AI as a smart two-level filter:

  1. Automatic Resolution (L1): Answer stock, hours, and location questions without human intervention.
  2. "Co-pilot" Management (L2): For complex cases, AI gathers information, verifies the customer, and hands the chat to the human agent with all context ready, drastically reducing handling time.

80.7%

TOTAL COVERAGE

Out of 4,688 chats in Dec, 3,785 were immediately handled by AI (Queue elimination).

68%

"CO-PILOT" MANAGEMENT

AI processed and prepared 3,191 complex cases so the human only had to close them.

$2,058

Extra Sales (Dec)

ROI + 230%

The agent closed 25 direct sales in chat, more than covering their monthly cost plus LLM.

Results: Financial Efficiency with GPT 4.1-mini

One key to success was choosing the GPT 4.1-mini model. This allowed ultra-low latency (instant responses) and minimal cost per conversation.

 

The financial result is clear: The agent is not an operational expense. With an average ticket of $82 USD, sales generated in chat paid for the technology almost three times over (Payback > 230%).

"The value isn’t just that AI responds, but how it sets the stage. During peak season, customers arrive anxious. AI serves them in seconds, finds their order, and gives us the context pre-digested. That let us survive the peak without doubling the payroll."

Daniella Barquero Elizondo

Customer Support and Digital Channels Leader

"The challenge wasn’t just answering chats, it was integrating a complex ecosystem. Connecting the agent to our ERP (CIMO) and online stores (WooCommerce/Shopify) in real time turned the bot into a real extension of our inventory."

Oscar Reyes

IT Manager at TEC

Ready to scale your E-commerce?

Stop spending time answering "Where is my order?" Hire a Zelty Agent to sell for you.

Schedule a 15-min Demo