Artificial Intelligence · Customer Experience
An informational chatbot for EDEL
Design of a Telegram chatbot that answers frequently asked questions about NFC business cards, prices, shipping and payment methods for a small business.
- Programme
- Artificial Intelligence · Master’s in Business Intelligence and Data Science, UEES
- Format
- Academic group project · team of five
- Period
- October 2024
- Botsonic / Writesonic
- Telegram
- Miro
Overview
EDEL sells NFC business cards and card holders. Because NFC cards are a relatively new and little-known product, potential customers have many questions about how they work, what options exist and how to buy them — and the purchase decision depends on getting those answers.
The project combined a theoretical part — what chatbots are, how they use natural language processing and how supervised and unsupervised learning apply to them — with the design and testing of an informational chatbot for the business.
The business problem
Answering the same product questions one by one takes time, and unanswered doubts can stop a sale. The goal was to give customers clear, consistent and courteous answers to frequent questions at any time.
EDEL Bot
Telegram
- Customer: Hi, how does an NFC business card work?
- Bot: Hello, and thank you for contacting EDEL. An NFC card stores your contact details and shares them when it is held close to a compatible phone. Would you like to see the card types and colours we offer?
- Customer: Yes, please. And how would I order for my whole company?
- Bot: Certainly. For larger orders we offer corporate packs. I can also tell you about delivery and pick-up options, and the payment methods available.
Objectives
- 01
Provide clear, accurate information on EDEL’s cards and card holders, including features, prices and functionality.
- 02
Suggest suitable products according to each customer’s needs.
- 03
Keep a formal, cordial tone throughout the conversation.
Methodology
- 01
Define the purpose
An informational bot focused on frequently asked questions, with no transactional functions.
- 02
Build the knowledge base
A document covering three areas — products (how NFC works, card types, colours, card holders, corporate packs), shipping and pick-up options, and payment methods — mapped in Miro.
- 03
Organise FAQs and configure answers
The phrases the bot should respond to were defined and answers configured in Botsonic (Writesonic).
- 04
Publish and test
Deployed on Telegram and tested by simulating a conversation with a potential customer asking about products, features, prices and the catalogue.
Key findings
- Telegram was chosen as the first channel for its ease of integration and use; WhatsApp or Instagram were noted as possible future channels.
- The bot was deliberately informational: it had no integrations with CRM, inventory or other external systems. These were identified as possible next steps.
Tools used
- Botsonic / Writesonic
- Telegram
- Miro
What this project demonstrates
- 1
Thinking about customer experience first: what a buyer needs to know before deciding, and in what tone.
- 2
Structuring business knowledge so that an AI tool can use it reliably.
- 3
Defining a realistic scope and being explicit about what a solution does not do.