How Data Kingdom stays ahead with the latest BI trends: ChatGPT
Important BI tools
Within our field, two important tools have been launched in the past year: Fabric and ChatGPT. In this two-part post, we would like to share what we as a company are already doing with these tools, how they help us, and consequently, how they benefit our customers.
Generative AI-tool
ChatGPT, a generative AI tool, is primarily used by us as a “Google on steroids” and as a source of inspiration for all sorts of things. We also use it as a tool to develop code and solutions. Think of DAX, Power Query, a step-by-step guide to setting up Power Automate Flows, or translating queries into Python, C#, and cURL.
Optimisation from ChatGPT-responses
With a few good prompts, an explanation of the situation, the goal, and a description of the metadata, we usually get results that our consultants can work with. More complex situations often do not produce the correct answer immediately, but they do provide inspiration and potential solutions. We then refine these in our products until we achieve the desired result.
Because usually, the final answer from ChatGPT, especially when it comes to code, is not yet good enough. It still generates errors, lacks nuances, and often misses those crucial details. For example, it might return a nice piece of code, but it cannot be refreshed from the cloud, or the figures that form the output are not always correct.
Current boundaries of AI
And that, I believe, is the crux of the current state of AI solutions. It helps, but as far as we’re concerned, you can’t completely rely on it yet. As Picnic’s CTO, Daniel Gebler, said, “we’re not going to order ten thousand bananas just because AI thinks it’s a good idea,” and the same applies to us. It’s a great tool, but not something to blindly trust.
However, it does help us. It makes us smarter, and the inspiration it provides generally improves our final results. We strongly believe that you can achieve more together than alone, and ChatGPT can, to some extent, fulfill the role that a second person would have in a ‘collaborative’ situation.
How this will develop further in the future is going to be interesting. AI techniques are improving rapidly, and we’re certainly keeping a close eye on it!
Steven Annegarn
King Expert Officer @ Data Kingdom
Please contact us!
