Prototyping in a GPT Hackathon
Defining product ideas based on GPT
Within a few months, ChatGPT has made us all aware that Generative AI will turn our professional world upside down. When and how this will happen is still unclear, that it will happen is certain. But what does it mean for an individual company? What are the meaningful use cases that can already be implemented? Where can the biggest efficiency gains be expected with GPT, and what does it take to realize them? In a hackathon, we worked out answers together with our customer within a minimum amount of time.
The challenge
In a March 2023 study, Goldman Sachs calculated that Generative AI can automate 25% of current jobs in the United States and Europe. At 44%, the number is particularly high in the legal sector. For our law firm clients, this means they need to increase their efficiency with AI in the short term and change their business model in the medium term if they want to remain among the leading law firms in Germany. But what is the best way to get started on this topic, and how do you get a process of AI innovation going?
The solution
Under the motto "Acquire knowledge, master methods, try it out", we held a hackathon with SKW Schwarz lawyers. The first item on the agenda was all about learning: What is Generative AI, anyway? And why will it turn everything upside down, as opposed to classical AI? At that time, best practices were still scarce. But with some interesting examples, we were able to kick off the discussion about the best uses for the firm. As an interim result, we had three promising product ideas on the table.
Together with our UX/UI designers and developers, we refined the rough concepts into concrete product ideas in small groups through a structured ideation process. The main focus was to turn the initial idea into a visible and presentable preliminary product.
We then presented the results to the large group. All three products were well received - in further exchanges we were able to gather more ideas for potential further development. After a small makeover, the prototypes were also shown to the colleagues who were unable to participate in the hackathon.
The result
In just one day, our client was not only able to develop a clear view of how important language models are for the future of their law firm. They also applied methods to move from vague ideas to a testable product. As a result, they were able to come up with concrete concepts that can give them an immediate boost in efficiency. They also found exciting approaches to building new revenue streams for the firm and putting them into practice.