dicobot - Behind the scenes of our AI chatbot

I'm Dicobot, the AI chat assistant from diconium. Currently, I am still learning how to talk to humans other than my developers but keep an eye out. I will soon be ready to make your visits to diconium's website more fun, direct, and dynamic. My superpower is to transform your questions into dico-language and give you the best possible dico-answer, while making suggestions about what you could be interested in based on what you’re asking. Want to try for yourself? Tune in around fall and say hi to me.

My development is a team effort. I think it's pretty amazing because my team gathers lots of insights from different areas, which helps them think about various aspects while developing me. We've got a Scrum team that covers User Experience, Legal, and Backend areas. This team collaborates in sprint-based cycles. You'll find amazing people like Joao Alexandre, Kristina Bicaj, Amin Dadashi, Hüseyin Demirci, Angela Hirsch, Eric McDermott, Joao Mello, Francesko Molla, Selman Özen, Mahdi Sadeghi, Juan Sehanues, Tiago Valente, and Talha Watto making magic happen.

During my alpha phase, my development team created me as a functional chatbot focused on providing you with helpful answers to your questions based on diconium-related documents. These documents are gathered using a custom web scraping routine, enabling me to sift through diconium's website content to pinpoint relevant information for creating personalized responses. To ensure I'm on point, a specialized retriever technology was also developed, making sure my answers align well with your inquiries. Additionally, I make use of the latest advancements in Large Language Model Concepts (LLMs) to better understand and address your questions. The team has diligently tested and refined my capabilities within a controlled environment to make sure I'm at my best.

Moving on to the beta phase, my development team is focused on refining my chatbot model and improving the user experience. They're setting up a secure testing environment where diconium employees can access and interact with me as a chatbot. During this phase, they'll be asking me various questions and assigning tasks, and their valuable feedback will be like guiding lights for the development process. They can mark each question-answer pair with a "thumbs up" or "thumbs down" and provide specific feedback, which will be used to fine-tune my model further.

In the Go-Live phase, I'll be accessible to all users on the diconium website. As a chatbot, I'll be smoothly integrated into the website experience, aiming to provide helpful answers to any inquiries about diconium and its offerings.

AI chatbots are real multitaskers. They can handle a bunch of different questions without needing a rulebook for each one. Plus, they learn from chats, so they can handle even trickier stuff over time.

Imagine chatbots that remember what you like and how you talk. They use that info to give you answers that really hit the spot. It's like having your own personalized helper.

AI chatbots are all about growth. With machine learning, they keep getting smarter with each chat. That means the more you use them, the better they become, all on their own. Exciting, right?

#CreatingDicobot: Behind the scenes of our very own AI Chatbot

A story from its perspective

Dicobot, can you share the milestones of your development?

Sure thing, so I grow and evolve in three main phases

What are the key benefits of AI Chatbots?

Well, I would say the three main benefits are:

Got any questions?

Kai Eder