It has been a new opportunity to test how our company structure enables us to react and integrate new technologies that offer real digital progress.

If artificial intelligence is nothing new, the public availability of tools, including the famous ChatGPT, has made AI even more indispensable. As avid geeks, we have been following the developments of AI and LLMs for years, quickly testing these tools in their beta versions. Like the rest of the world, we were impressed by the performance of ChatGPT and others. Sometimes overwhelmed by the possibilities unfolding before us.

How our self-organisation allows us to evolve with new technologies

In a self-organisation like ours, innovation is not the responsibility of a closed department or the result of a top-down impulse. Instead, it is the result of individual and collective actions that allow us to evolve and organically transform our structure. These technologies raise questions, including their alignment with our values. In other words, is AI a true digital progress and not just an innovation?
Our AI experts from various circles and backgrounds quickly emerged. They took initiatives, especially in the education sector, collaborating with our training experts. In a few weeks, AI training sessions were offered to all Liipers, including specific modules for developers. Our annual LiipConf included several workshops on AI, and several Liip Talks and blogs were also dedicated to the topic.

Purpose over Profit

Apart from the aspects of creating text or image content that garnered the most attention, AI tools also enable code generation. As a web and mobile development agency, this naturally caught our interest. The question of being replaced by machines has become relevant in recent months, rightfully so in many cases.

As for our development work, we appreciate being able to use ChatGPT to generate simple and repetitive code. This allows us to focus on more complex coding tasks beyond the reach of AI tools at the moment. We are fortunate to operate in a professional environment where the goal is not to maximise profit and profitability at all costs. This evolution, if not revolution, in our way of working does not involve layoffs but rather an evolution of our organisation. However, we are aware that this is not the case in other companies and industries.

New AI technologies have made headlines, not only for good reasons, with the issue of related layoffs not being the only concern. Behind its impressive technological facade, a tool like ChatGPT also raises environmental questions, with significant energy consumption in its data centres, as well as ethical ones as well. We could easily sidestep the issue since these are not our own tools. Nevertheless, it is crucial for us not to use ChatGPT just for the sake of using it and staying up to date. On the contrary, our goal is to identify contexts where its use signifies real digital progress. We do not necessarily provide a global answer to the question of AI as progress or not. However, we do identify scenarios where AI can solve real problems.

Practice over Theory

In addition to training, our education budget allows us to learn on the job, which is especially important for emerging technologies. True to our principle of practice over theory, the ZurichGPT project was born.

Who has not wasted time searching for information on a municipal or cantonal administration's website? Recognising that many public pieces of information are unfortunately hard to access, Christian Stocker developed ZurichGPT, an AI-based chatbot. Rather than having to search for answers on the general web, our tool bases its responses on a unique corpus, the website of the city of Zurich (and its PDFs), reducing the chances of "hallucinating." One can ask ZurichGPT questions in their preferred language, even if not supported by the site, to get a sourced answer. This is real progress in terms of accessibility of public information.

ZurichGPT is not affiliated with the city of Zurich, but they appreciated the initiative. After this initial feasibility study, the platform gave rise to LausanneGPT and GenevaGPT to refine the tool. From this practical experience, we were able to benefit the practical legal information site lex4you with lex4youGPT, a bot based on GPT. Thanks to it, anyone can ask legal questions and receive a specific and documented answer based on the rich information on the site.

LiipGPTs are just the beginning

Lex4you is our first client project based on generative artificial intelligence, and this is only the beginning. These new developments enhance the accessibility of information search compared to writing prompts. We are also testing machine learning tools to improve the contextual understanding capabilities of local search engines. Furthermore, we are performing various experiments with our clients from government, education and industry to find potential in their field. This also in connection with image recognition, text to speech and speech to text. Again, the idea is to offer pragmatic solutions to real issues, not just AI solutions for the sake of it.

And we are happy to discuss with you the potential AI solutions that really fit your needs and strategy. Contact us.