ChatGTP

Everyone has heard of ChatGPT, and it's true: it's a powerful tool, capable of composing a letter, assisting with homework, and so much more. However, what a lot of people are missing is that its potential extends far beyond basic language generation requests. It would appear that ChatGPT is the first truly versatile, all-purpose general AI.

At www.nexoid.com, we have been experimenting, and found ChatGTP is exceptionally good at abstract categorization. Consider this scenario: you're a help desk manager, and you're tasked with evaluating your team's performance. How can you discern which agents are excelling, and who might be struggling? As it stands you could measure how many tickets each agent closes over the day, but we all know that person who cleverly cherry-picks the easiest tasks.

Now imagine a system where an AI, like ChatGPT, is capable of reading a support ticket and accurately estimating not only how long an agent will take to resolve. This single datapoint is a game changer.

But before you rejoice, ChatGPT's abilities extend way beyond time estimation. It can be employed to assess security risk, determine the wider impact on the company, identify the underlying technologies or assets involved, and even designate the most suitable team to handle the ticket.

We have now entered a world where AI is capable of not just proving better management information but automating some human tasks.

One of our many experiments was using ChatGPT as a chatbot. It turns out it is possible for ChatGPT to manage a three way conversion acting as the interpreter between a user and an API. Traditional chatbots operate by developers mapping out user intentions. This approach's effectiveness depends heavily on the developer's foresight. If business logic is missing or cases become more complex, it could lead to frustrating experiences for end-users.

ChatGPT is designed in such a way that it can be taught much like a human. By providing it with a list of APIs or functions it can access, ChatGPT can independently derive the necessary business logic. When we first saw this and realised what it was doing we were amazed. Even heiring intelligent experienced people it is often weeks before new staff understand and fully appreciate the Nexoid API.

By eliminating the need for developers to anticipate user needs and instead relying on the AI's raw intelligence, we've observed a significant increase in both user satisfaction and problem-solving efficiency.