We fully agree with the perspective that Agentic AI is only as intelligent as its entities, and entity resolution is a key 'track' for any 'Formula 1 car.' In our experience, the problem lies not with the agents themselves, but with the architecture that underlies knowledge management. We would be interested in whether you are exploring solutions that are radically different from those based on vectors, which, as research shows, have inherent limitations in terms of structure and provenance. We believe that it is precisely in these different foundations, such as hypergraphs, that the potential to overcome the chaos of entities and achieve truly reliable results lies.
We fully agree with the perspective that Agentic AI is only as intelligent as its entities, and entity resolution is a key 'track' for any 'Formula 1 car.' In our experience, the problem lies not with the agents themselves, but with the architecture that underlies knowledge management. We would be interested in whether you are exploring solutions that are radically different from those based on vectors, which, as research shows, have inherent limitations in terms of structure and provenance. We believe that it is precisely in these different foundations, such as hypergraphs, that the potential to overcome the chaos of entities and achieve truly reliable results lies.
Yes I agree that we need the right foundations in data management. Check Zingg for an opinionated approach