CONTRIBUTED BY
Karolina, ExpertHub Team
DATE
Oct 27, 2025
Last week, we had a great pleasure to not only participate but also speak at a conference in Poznań. Karolina, our Chief of Staff, shared our company’s goal of responding effectively to the AI adoption crisis with our new platform, SolveStack.ai. The Entrepreneurship Days attracted audiences and speakers from across disciplines, but Wednesday, 22 October, was all about AI.
Karolina offered some insights into the startup vetting process, presenting examples of interesting AI applications and describing the factors that make AI startups attractive to the market. She also talked about the role of community development in startup work, demonstrating some unique paths of human engagement in AI projects.
The chance to show our perspective was great, but the learning opportunity presented by the conference was perhaps even more important. We found the other talks inspirational, and in this blog post, we want to share with you the most crucial take-aways.
Grow with AI: From Vision to Product
Last year, the Conference introduced the topic of artificial intelligence for the first time, from a very abstract and theoretical point of view. This edition, as the average understanding of the field evolved enormously, organizers emphasized AI in practice. Paradoxically, with the development of AI, the role of humans has begun to matter more and more.
During the event, we heard about real-life AI implementation examples from both startup founders and representatives of large enterprises. At the end of the event, we all collectively came to the conclusion that AI will not replace humans, but it will replace humans who don’t know how to work with AI.
Data is the most expensive currency
Since 2021, the investments in AI have tripled. But the ROI is still unsatisfactory, with 70% projects failing to deliver it. The first Conference speech by Maciej Szczerba from SherpaSearch tackled the theme of investment markets being now more competitive than ever for AI startups. Innovators ask themselves continuously: how can we stand out? One of the firmest answers is: data.
Data drives the willingness of implementing AI in large scale businesses. Having data means having power. Metadata access, data-labelling technologies, and capacity to build one’s own LLM are however often out of reach for startups.
This is when the business skillset comes to the fore. In times, where prototypes are built quickly and cheaply, one has to prove that the product is durable, marketable, and profitable. And it happens through the right partnership outreach, marketing, and pitching. “The winner is the one who understands the human (needs), not the machine”.
On top of that, the product or service should be built with compliance in mind. Compliance, as Szczerba described it, means understanding what data mean in a specific context.
AI is more than you think
Joanna Bigaj from Allegro demonstrated with passion how AI can be both useful in improving company’s products and productivity. Simply put, it can be operationalized not only to provide benefits for customers but also to augment the team workflows and personal performance at work.
What was particularly striking was to learn that Allegro has been utilizing AI in its offer for over 12 years now, making us wonder about the reasons behind today’s AI hype. How did they do it? One example would be the intelligent search engines, able to learn based on customer buying patterns or merchants’ styles of product description.
Joanna’s stand on human participation in the loop of AI project related, inter alia, to interpreting obtained data. “Correlation between trends does not mean trends stem from each other, and therefore, a broader orientation in the market and data interpretation is needed to work with AI effectively”.
AI & Law
As artificial intelligence keeps spanning the tech fields, it begins to do so also in the industries traditionally not associated with technology at all. One of these areas is law. Professor Michał Jackowski demonstrated why and how AI can be a remedy for tedious work of lawyers, allowing them to focus on more critical operations within their practice, to "do the actual thinking".
Artificial intelligence tools have the right potential to reduce a lot of time that for years has been wasted on administrative tasks and bureaucracy in this discipline. Jackowski proposed a thesis that AI and lawyers will never be competitors but rather may become partners. Moreover, AI creates new areas that people would need legal assistance with. He noticed, however, two major loopholes.
Both law practitioners and their clients are still reluctant to implement AI, viewing the field conservatively. Moreover, there are many ethical questions surfacing, regarding trust, confidentiality, reputation, data protection, and such. It is often easier to continue the traditional way instead of challenging oneself to implement a new tool that, in the long run, can be way more effective.
The understanding of legal tech framework is poor. The awareness is limited to the existence of ChatGPT, while knowledge of fine-tuned models, adjusted to specific tasks in a specific area of work is scarce. This translates to general reluctance in AI adoption.
It was Michał, who said straightforwardly: AI will replace precisely those lawyers who do not know how to use AI.
AI in Enterprise vs AI in Startups
The last two panels demonstrated perspectives on the AI project pros & cons. Krzysztof Sopyła shared his take on implementing AI in a larger enterprise (Pearsons) while Ewelina Skowron offered her startup-derived insights (youbloom).
The use case Krzysztof based his presentation on was Digital Language Tutor – AI assistant for learning languages. His presentation was a story of building a product with AI and convincing the company that it was a good idea. Sopyła demonstrated interesting data visualization methods used for moving the project forward, work distribution strategies across a diverse but relatively small team, where some people needed to play multiple roles, and guided us through the opportunity the whole initiative brought for the company.
There are also challenges he experienced that many companies we address with SolveStack.ai can relate to. Misalignment between teams, lack of leadership support, time wasted on pitching, explaining, and finding the right specialists… These difficulties restrain AI adoption development globally.
Ewelina, as a startup founder, had faced different obstacles. In startups, there is a lot of ideas, willingness to innovate, explore, discover, and build. AI enhances this creativity, opens new possibilities and offers avenue to build your services and products with it. Such technology can also simply improve the quality work of startups, where there are often not enough hands on deck, and a general lack of resources is daily reality.
However, rapid AI solutions development can also induce stress, as Ewelina mentioned. The chaos and abundance of artificial intelligence tools, the need to stay up to date, and the peer pressure cause some organizations to lose time on looking for the right solutions.
Knowing these challenges has made us more ready to develop our platform SolveStack.ai in the direction that will support both enterprises and startups.
Thank you for the opportunity to be there, Poznań and see you next year, hopefully!
As our team continues adventures with conferences in Europe, stay tuned for updates related to SLUSH in Helsinki this November.
