Automated experimentation platforms in polymer chemistry have greatly amplified the number of experiments that may be performed and the breadth of polymer composition space that may be accessed. These rich experimental datasets can be utilized for development of new AI models or agents that can accelerate traditional research workflows via hypothesis generation or reactor process control. However, experimental data from polymer synthesis and characterization is inherently complex and spans multiple domains ranging from organic synthesis to catalysis to polymer formulation and processing. This requires an inherently highly flexible and extensible data model for accurate representation within database systems and translation for consumption by AI development pipelines. To facilitate the connection between AI development and automated experimentation, we have developed an experimental polymer database system. Here, we discuss our approach to addressing complexities of handling experimental data as well as the structural representation of stochastic polymeric structures—a critical feature to allow interoperability with other datasets. Integration of this system with broader efforts in the chemistry community in order to facilitate collaborative, cross-lab research on polymer informatics and using AI and data science to accelerate materials development will also be discussed.