ModMed has been offering their cloud-based EHR for 15 years, so they have built up a database on more than 600 million unstructured texts in real-life clinical use. Therefore, now that they are adopting AI, their ambient listening tool can generate specific, highly accurate structured data from a patient interview. They don’t need a Large Language Model (LLM) the same way other ambient clinical voice companies need them, according to Cofounder and Co-CEO Dan Cane, so they can avoid the errors and lack of transparency found in LLM-based solutions.