 How you identify the most valuable data, how you can manage the right dataset for the AI project will be the biggest challenge. Data visualization, a visual depiction of data, a graph, a map, an infographic. They are not just images, they are not just pictures, they are visual arguments. Therefore, they need your attention if you want to understand them correctly. Understanding what's possible, what's desirable, and also understanding what's just part of the hype curve of AI. They lack from having trusted data, which is not just about cleaning data, it's something more than that. It's data privacy, it's data security, it's data compliant. Your average company at least here in Spain doesn't have enough digitalized information to make sensible decisions. Giving a step from models to production and to being able to effectively use those models. Just even going into the companies and getting the data infrastructure in place so that a data science team can work with other departments and the data is not siloed. That is a fundamental thing and maybe only half the companies have even started this. Keeping it simple, breaking apart problems into smaller bits. Cleaning data is absolutely mandatory if you're going to have high quality machine learning. Really building products more than just doing projects is the biggest one. Communication to everybody about this world is really important. I think for me the challenge that I'm exposed to most is actually the underlying data itself. Where's one? Change the data, transform the data to trust the data. Second one, use AI. Make the AI models and integrate those models inside your business processes, your business applications, your businesses. It's around the regulation because it's really hard to be able to get valuable data. It's really important to have the most proper team that can extract all the value from this data. It's time now to think also about the CO2 footprint of artificial intelligence.