The same AI models are accessible to basically everyone. So why will some organisations win while others won't? (podcast recap)