Blog Data & AI Modernization

How to boost your productivity with AI-infused solutions!

AI is becoming a more important part of the way we work. It can help with creativity, productivity, automation, data analytics, and more! 

At Intercept, we increasingly use AI to improve our solutions, for ourselves and for our customers. In this article we dive into AI infused solutions that help to boost productivity. Hopefully it gives you some inspiration to further implement AI in your way of working. Let's dive right into it!

Author

Kaijisse Waaijer Azure Data Scientist

Reading time 3 minutes. Published: 16 May 2024

AI Copilot tools

Several Copilot tools have been released in the past year, such as Microsoft Copilot Suit, Copilot Studio, and the additional features for GitHub Copilot. In the past months, at Intercept, we agreed that the Copilot tools mentioned boost our productivity and creativity. They are all increasingly becoming part of the toolset that we use every day.

Let’s take GitHub Copilot as an example; we are infusing our Integrated Development Environment (such as Visual Studio Code) with AI, helping us generate code and focus on business logic and specifics instead of typing large chunks of code. But Copilot is not the endpoint. We are just scratching the surface of what AI can do.

Microsoft Azure AI tools

With Azure OpenAI Assistants, we can already create our own AI assistant. With readily available code examples, we can have one up and running in minutes. By adding our own data sets and leveraging features such as Azure AI Search, we can even use the generative power of large language models (LLMs) such as ChatGPT and create valuable responses based on our own data.

But let’s take it a step further. Think about repetitive and time-consuming tasks. We are going to examine business processes and process automation…

AI infusion in ITSM

Let’s take the following example: you have an IT Service Management Process and a solution that allows your customers to create support requests. Imagine connecting your existing product documentation, support process documentation, and even sales collateral to your Azure OpenAI Service Instance.

This is where the power of Azure AI Search comes in. By extending the knowledge of the LLM in Azure OpenAI Service with your own data, you can then respond to a customer based on your own document sets and leverage the generative power of the LLM in Azure OpenAI Service.

Note that some prompt engineering is involved in setting the right context, but the concept is very promising. How cool would it be to have the following happen:

Customer: “I would like to change the dashboard layout you provide, but I don’t seem to be able to do that.”
AI-Generated response: “Changing the dashboard layout requires ‘Edit’ permissions and can be provided by the dashboard owner. You can identify the dashboard owner by clicking the ‘Info’ button at the top right of the dashboard view.”

That does look a bit like the classic chatbot. However, as we are using an ITSM system, communication takes place.

An example like this is just the beginning. We can take it even further and start automating processes. We could use AI to screen a ticket that comes in automatically. And if there’s a valid feature request, it automatically creates that request on your product backlog.

We are using pretty basic but powerful examples here. All of this is already possible (based on the quality of your dataset and standardized business processes).

Kubernetes AI Toolchain Operator (KAITO)

Azure OpenAI Service provides a limited set of language and image models. But what if you want to leverage generative AI to process some financial data? The models currently available in Azure OpenAI Service are not suitable for that.

We could use a custom model from the AI community's Hugging Face. However, that often requires some computing and can be expensive compared to the usage-based model of Azure OpenAI Service.

If we look at the example of processing financial data, that is typically not something you would do 24/7. You just need the model to be available when you want to process the data. This is when popular platforms such as Kubernetes come in.

Recently, Microsoft released the Kubernetes AI Toolchain Operator (KAITO), which allows us to provision infrastructure (GPUs) when we need it. We can even delete the expensive GPUs/compute afterward, only paying for the hours that it runs. This makes KAITO suitable for scenarios beyond Azure OpenAI Service for affordable pricing.

AI-infused solutions are the future

We can infuse our solutions with AI and improve efficiency, productivity, creativity, business intelligence, customer experience, and more. Microsoft has recently released many new AI tools that make our work easier and more interesting! We are all excited for what is to come.

Lara Lamberts

You are still early!

Become a pioneer within your industry by leveraging the power of AI in your solution(s).