Yes. David, this is Johnny. Very good question. Thank you for that. Yes. So, we're excited about where we are with Clarity today. The product is in full production, and maybe to give a little bit of insight of what the product does, it applies artificial intelligence and the underlying LLM models to extract or to understand basically unstructured documents, primarily for us, obviously, coming from the fax world, that is also capable of ingesting documents from other channels. And then extract based on the prompts that were programmed individual data points and put those into a structured data format and then import those in, for example, an EHR system or any kind of system, through a variety of file formats, we can customize those for our customers. So, there's a lot of demand for that. There's a lot of POCs. We have a few systems actually in production already. So, we're able to win customers to sign contracts. We're in deployment, rent production. There's a lot of ideas of what can be done with the technology like that, which makes it a little bit more difficult on the deployment side, which is why we need a lot of proof of concepts to really fine-tune this for individual customers. But we're also very much focusing on developing replicable solutions. We've mentioned this in the past, we have two models that are out there right now, one for prior authorization purposes. So, understanding of prior authorization requests extracting data points there and accelerating the processing of those? And secondly, basically categorizing and classifying clinical documents and then extracting primarily patient demographics but also additional data points regarding insurance and those kind of things and then creating a CCDA document and then forwarding that document in a structured format, and basically providing two things, the unstructured documents it can still be read by the human eye plus a structured data file with these demographics. So, that accelerates the process of filing these documents within, for example, EHR systems. What we're currently not doing is really extracting things like diagnosis or more complicated text. We're really very much focused right now on the structured data fields.