You know ChatGPT, right? Everyone does. For us at SolDevelo this tool was a game changer – it helped us with many projects, most substantially with OpenLMIS. In this case study you will learn how exactly AI is helping us with our daily tasks, why QA with ChatGPT is such a powerful combination, and in what ways it supports us in improving OpenLMIS.

OpenLMIS

OpenLMIS is an open source platform tasked with facilitating medication management across hospitals and other health facilities, which means tracking medication orders, managing physical inventory, monitoring for expired items and many more. Additionally, it has other features like cold chain equipment tracking, as some medications are not shelf stable in room temperature, extensive reporting features, or certain offline capabilities to allow work during unexpected internet interruptions.

The Challenge

While the OpenLMIS application itself doesn’t sound exactly simple, there’s another curveball – our team takes care not only of the core application itself but also certain country-level implementations of the platform, providing development, support, and Quality Assurance to each of these. As it usually goes, every implementation has its quirks and additional features, which ultimately makes for a quite challenging project.

Due to being responsible for the core application, as well as various implementations, the knowledge we need to process can be best described as fragmented. Multiple implementations have their own Confluence pages, so that only relevant information is contained there. Unfortunately, this makes it difficult to consolidate all the available knowledge, therefore we end up with multiple separate sources. That, combined with the myriad of features available in the application, equals a situation where verifying against documentation can be challenging.

The Solution

Even though challenges are inevitable in any process, especially in complex fields like Quality Assurance, having the right tools can make a significant difference. By leveraging advanced technologies such as AI-driven solutions, we can streamline laborious tasks, and increase our efficiency.

Smart tools for smart results

This is where ChatGPT comes into play to aid us with managing all that. For as long as LLMs (Large Language Models) existed, which is still not that long, if you think about it, they’ve been useful for assisting developers with coding and debugging. Meanwhile, on the QA side, it’s quite often used to generate test data and help with more menial tasks such as creating test cases, or producing all sorts of documentation. These use cases are definitely important for us, but the groundbreaking feature that we were waiting for has dropped in ChatGPT 4, and that’s file upload. 

The complexity of the system that we have on our hands is daunting to say the least. We have a lot of documentation in many different forms – spreadsheets, Confluence pages, text documents. Getting to know all this information is very time-consuming, and there’s no guarantee that after reading all these documents we’ll remember even half of it. 

New possibilities with ChatGPT 4

So, what if there would be someone who can remember everything flawlessly and be able to help everyone when they need a double check on anything system-related? Well, that’s ChatGPT. While it’s not perfect, there’s no doubt that retaining knowledge and being able to summarize and explain complex system matters are the skills it excels in. 

Of course, having this tool to our advantage doesn’t mean that we don’t have to be proficient in how the system works. It’s actually the other way around – we’re more proficient with the system because we have a source of easily accessible knowledge that we can trust. Additionally, being able to upload the files also enhances ChatGPT’s context to help us with the previously mentioned tasks, making its responses more accurate, further reducing the time needed to verify and adjust its outputs to meet the criteria.

Keeping a finger on the pulse

We’re always on a lookout for new tools and technologies to boost our productivity and quality. Going with ChatGPT was definitely a win for our team and we can’t imagine working without it. 

Personally, I have a feeling that we’re only at the beginning of the AI transformation, so I’m looking forward to what the future holds in terms of further improvements in project management and delivery.