AI trends and industry changes for 2025: An executive view

When evaluating AI tools, it’s helpful to focus on specific use cases. Fall risk is a good example. Often, changes in daily activities or medications, a new diagnosis, or even something as basic as dehydration can indicate an increased risk of a resident falling. If this information is gathered and given to the right person in time to affect a patient’s care, it can make a big difference.

De-identified data in senior living settings is showing us a strong correlation between a person’s changes in health and an increased fall risk. When caregivers can identify and proactively manage behavioral changes like sleep habits, cognitive abilities and speech patterns, resident outcomes can be significantly improved.

There has also been a lot of talk in the past few years about remote patient monitoring (RPM)  technology, including everything from wearables and home-based equipment to passive monitoring of vital signs and other health metrics. I’ve seen mixed results from these tools. Some studies show improved outcomes and reduced hospitalizations, and others showed little correlation between RPM solutions and better outcomes. Combined with issues managing equipment, training users and covering costs of RPM devices, successes were hard to find.

More recently, there’s growing interest in passive monitoring, such as LIDAR (Light Detection and Ranging). These systems can track changes in a patient’s activities of daily living as well as events like fall detection, and correlate that data to changes in a patient’s health—both improvements and declines. Tools like LIDAR are a great example of the three Ps mentioned above: passive, predictive and proactive.

We’re also seeing work on using AI to create personalized care plans. AI can evaluate a patient’s medical history, genetic dispositions and lifestyle changes to develop care plans. They can also review any history relevant to a home health or hospice episode—for example, a knee replacement or COPD exacerbations—to develop a care plan that will keep that patient out of the hospital and make sure they have the services they need.

Another area of active research is using AI-powered chatbots. Chatbots have been shown to deliver more accurate information when answering patient questions on several topics, and have the advantage of being available 24/7. To help both clinicians and patients become comfortable with chatbots, I recommend starting with simple use cases. To ease clinicians and patients into using chatbots, I suggest beginning with straightforward applications. For instance, during a home visit, a nurse could use a voice-activated chatbot to quickly check the next therapist appointment instead of relying on a tablet or mobile device. This approach highlights the tool’s simplicity and accuracy in real-world settings. Chatbots can support the patients and family caregivers, as well, by providing basic information such as when the next nurse visit is scheduled, rather than calling the home health agency or texting the nurse.