ALIS Innovation, Technology & Agility

How is your team integrating AI and ML into the product development process, and what specific improvements have you seen as a result?

Over the past five years at Medtelligent, we have developed and refined a pipeline of taking ML/AI solutions to production which starts with identifying pain points that may be amenable to data-driven solutions. In doing so, we prioritize rapid iteration over prolonged analysis. This helps us avoid investing significant time and resources in solutions that may not eventually be successful in production.

Once the model is finally in production, we don’t simply set it and forget it. As part of our productionization process, we have established a feedback loop that captures and analyzes model errors. These errors are then used to refine the model and improve its performance over time. As the model matures, we offer users flexibility in how they interact with it. Users can choose from a range of options, from using the model as an assistant to help with specific tasks to fully automating certain processes. 

What strategies are you employing to ensure that your systems and processes keep up with the rapid advancements in AI and ML?

The landscape of AI is constantly evolving, with new techniques, tools and applications emerging every day. To keep up with this rapid pace of change, any organization that looks to integrate this technology in their products or services must inevitably embrace a culture of continuous learning and adaptability. This sentiment is reflected in our company’s tagline — welcome change — which was intentionally chosen to embody this core value. What also helps is the autonomy the data science team has to set the AI/ML direction for the company, which makes it more agile in responding to change. On a tactical level, we hold weekly ideation sessions where we discuss the latest AI trends and tools and how we can use them to “make life better” for assisted living and memory care companies, communities, staff, residents and families.

Can you share some examples of how AI/ML has directly contributed to enhancing your product line or accelerating time-to-market?

There are too many applications to mention, but I will share a few examples. We’ve built an AI-powered early alert system that proactively identifies at-risk residents. By analyzing a diverse set of data points, ranging from clinical notes to medications, the system detects recent changes in a resident’s health status and uses this information to flag residents who may be at risk of departure from their communities. These early alerts allow staff to intervene and implement strategies to avert resident move-outs. 

We also leverage LLMs to process and extract insights from unstructured data. This includes free-form text data from observations, progress notes, medication records and incident reports. Besides text, images constitute another major type of unstructured data within our electronic health records. Today, the healthcare industry still heavily relies on paper-based documentation and record-keeping. Similar to using a mobile check deposit app, where a photo of a check is processed to identify the payee and amount, we utilize computer vision models to analyze scanned images of various forms and documents generated in the context of senior living.