Chatbots

Multimodal Customer Support Bot

Nexora Telecom

OpenAI API PHP MySQL
80% ticket deflection

The Challenge

Nexora Telecom's support centre was drowning in repetitive tier-1 queries — billing questions, SIM activations, data top-ups — while their agents spent 70% of their time on issues that required no human judgement.

80%
Ticket Deflection Rate
24/7
Availability
4.7★
CSAT Score

The Solution

We built a multimodal support agent powered by GPT-4 with retrieval-augmented generation (RAG) over Nexora's knowledge base. The bot understands natural language queries in both English and Albanian, maintains conversation context across sessions, and integrates directly with billing and provisioning APIs to take real actions — not just provide information.

Deployment spans web chat, WhatsApp Business API, and email via a unified PHP backend with MySQL conversation logging. Every interaction is scored for sentiment and intent drift, feeding a continuous improvement loop.

Complex or sensitive queries trigger a seamless handoff to a human agent with full conversation history, ensuring no context is lost.

OpenAI API PHP MySQL WhatsApp Business API Redis Elasticsearch

Key Results

80%
Ticket Deflection Rate
24/7
Availability
4.7★
CSAT Score
Project Screenshot 1
Project Screenshot 2
Project Screenshot 3

"Our support volume grew 40% last year, but our headcount didn't. That's because of Vixus's bot."

DK
Dorian Kelmendi
Head of Customer Experience, Nexora Telecom

Computer Vision QC System

AI Development

View Case Study

Ready to Build Something Like This?

Tell us your challenge. We'll design the solution.

Start a Project