Does AI Free Senior Living Staff for Human Touch? Only If you Plan Where the Time Goes.
TL;DR
AI frees staff for human contact only when operators decide in advance where saved time goes. Training is essential to this process.
Kaiser Permanente’s AI scribes saved an estimated 15,791 hours. 39% of patients said their doctor spent more time talking with them directly.
Keep people on emotional conversations, give AI predictable tasks, and measure the effect on residents.
AI can offer senior living staff more time to connect with residents and families. That said, AI only frees up staff time for more human contact when operators decide in advance where the saved hours go. Hospitality and health care have more experience than senior living, so their results can guide the way.
Time savings are real
Hotels have moved routine contact to software. This year, Hilton told an industry conference that messaging bots handle 40% to 45% of incoming guest messages, according to a consultant's summary of the event. Wyndham gets thousands of calls a day from guests who only want to confirm a reservation. Those are the tasks AI handles well: frequent, predictable, and low in emotion.
Senior living also has frequent, predictable, and low-emotion tasks. At Atria Senior Living, once residents could ask an in-room Alexa device questions about meals and activities, emergency pendant use fell 30% to 40% in some communities. Atria's CTO, Chris Nall, said staff told him the devices let them "focus more on the care of the residents". Phoenix Senior Living says AI transcription saves its sales team dozens of hours each week.
Health care measured the follow-through
The best evidence comes from medicine. The Permanente Medical Group gave AI scribes to 7,260 physicians, who used them in about 2.58 million patient visits from October 2023 to December 2024. Kaiser notes that the tool saved an estimated 15,791 hours of documentation. In patient surveys, 47% said their doctor looked at the computer less, and 39% said the doctor spent more time talking with them directly. One-third of patients, however, noticed no change.
A field experiment in Alibaba's customer service shows that when an AI agent took over only some chats, the chats people still handled ran shorter but earned slightly higher ratings. Staff did better human work once part of the load was lifted. The same study found the worst ratings when software handed an upset customer to a person too late.
Time can also leak away
Be intentional about the AI tool’s “why.” If the goal is to simply automate tasks without additional attention in your value proposition, you might be able to reduce headcount at best. If the goal is to automate mundane tasks so that staff can focus on your value proposition and engage more prospects and residents directly, you’ve started down the path of augmentation.
Hours saved don’t automatically translate into new productive uses. Isaac Scott, CEO of Anthem Memory Care, estimates AI gives him back six to eight hours a week, but his team also spends time double-checking AI errors. Rick Westermann, CMO of LCS, described early overuse that produced long, unreviewed documents and added work, aka “AI slop.”
Families are watching for the leak. In Gartner's survey of 5,728 customers, the top worry about AI in customer service was that it would make reaching a person harder. AARP found that 68% of the adults it surveyed worry AI will reduce human interaction.
Train staff to adjust their workflows when AI tools are deployed. Argentum reports that nearly three-in-four workers already use AI tools. The report suggests awareness campaigns are far less useful than competency programs. Awareness can’t be tied to job responsibilities or resident experience. Competencies are observable skills that can be tied to roles and responsibilities. Teach staff to check AI outputs before they reach a family, prospect, or resident, and to recognize when emotion enters a conversation and needs a person.
Putting AI in senior living into practice
In more than 250 mystery shops of senior living communities in 2023, Bild & Co. found that 53% of web inquiries got no response within two hours and 13% of calls never got a response. That's an obvious place to send recovered time.
Consider these four practices:
Identify the tasks and conversations that carry emotion from the first minute. Keep people on these conversations and train them to intervene early instead of waiting on a handoff.
Give AI the predictable tasks, such as menu questions, reminders, and call notes. Train staff to review what AI produces before residents and families see it.
Name where time savings go. Options could include response time to families, minutes of resident contact per shift, or event attendance. Train staff for that work.
Measure the impact on residents, not just hours saved for staff. Kaiser demonstrated results by asking patients.
An operator that can't say where the freed time went has no evidence it reached residents.
Dan Lindberg is the founder and principal of Applied Economic Insight ® LLC.
Comments