Senior Living Operators Find Good And Bad In New Ai Platforms

It’s been more than two years since senior living operators introduced AI into their workflows. It’s been a mixed bag of results so far.

Operators have implemented AI into sales team practices with processes that include taking notes, roleplaying and analyzing leads. In other instances, it helps streamline operations at the executive level, which is the case for West Linn, Oregon-based Anthem Memory Care. Leadership teams are getting the most bang for the buck when it comes to analyzing reports and pulling out important information, according to CEO Isaac Scott.

The time savings as a form of measuring the return on investment from the executive team comes out to roughly four times the amount invested in AI platforms such as Claude and ChatGPT, Scott said.

“I don’t tend to dwell on the dollar amount, but in a given week I may be getting six to eight hours back that I didn’t have before,” Scott told Senior Housing News.

Not every operator has had perfect experiences with AI. For example, automated platforms don’t always take the best meeting notes and that can muddy instead of clarify post-meeting summaries, according to LCS Chief Marketing Officer Rick Westermann.

“I think there was all the hype around AI is going to fix everything, but you have to have your organization set up to enable that, and that takes a lot,” Westermann said. “You have to have that foundational layer … There are still systems that are within their own environments where we can’t deploy tools yet.”

Avoiding AI ‘slop’

Not all AI platforms have had a smooth senior living rollout.

Anthem encountered AI “hallucinations” that put out incorrect information, which hurt down productivity as staff double-checked notes and reports to ensure they were correct.

AI platforms that looked good on paper would sometimes provide conflicting information on the same data sets. The inconsistent results complicated the company’s process used to create pro formas.

“That was a lot of angst and effort that we put into a couple services that just didn’t end up panning out, so that was a disappointment,” he said. “Where it’s disappointing is when we’ve expected it to do everything, and been disappointed that it hasn’t done something.”

Anthem’s staff have tempered expectations about the realities of what the technology is fully capable of as a result.

Leaders with Centennial, Colorado-based Ascent Senior Living are still adopting an AI platform, according to Founder Tom Finley and Director of Business Intelligence Saranya Jambunathan. The operator’s teams already were building a data platform in 2021 before the rise of AI and large language models. The company is now beginning to use Microsoft Copilot to provide daily summaries to community leaders. Its teams are documenting everything so the AI has enough data to work properly, Jambunathan said.

Part of the process has been deliberately slow to put guardrails in place to protect Ascent’s sensitive data. Resident and staff data that could violate HIPAA cannot be uploaded to the platform and used to receive guidance, for example.

For LCS, Westermann said early on there were instances where AI was overleveraged to the point it was creating “slop.”

“People are finding these 10 page documents because they think it’s impressive, but it hasn’t been thoroughly reviewed, or it’s not on topic for what actually is going to solve the business case,” Westermann said. “That definitely creates some work for us outside of our normal work.”

Finding ROI on AI

Despite the challenges of adopting AI, operators have found ways to make it work for them.

One of LCS’ data scientists utilize Anthropic’s Claude platform to do research and help build prototypes of code infrastructure for additional machine learning, saving dozens of hours of work. The company’s sales team is using Salesforce AI to analyze prospect profiles and score leads.

LCS still is justifying the ROI for its 15 Claude Code licenses and Salesforce expenses.

“The cost is really inexpensive to do it in house, and so it’s something where we know we’d get some gains on productivity from helping our team,” Westermann said. “Ultimately we also have found out what we can garner from notes in the sales process that we can further incorporate into lead scoring.”

Ascent has, through its focus on building a data platform and using daily summaries on residents, increased length of stay in unit types such as assisted living, growing the average to 135 days over the past several years. Jambunathan added AI also supports caregivers by sending reminders about missing documentation to improve compliance while reducing administrative burden.

“There’s a lot of education that still has to happen, and that we believe AI can help us get there,” Finley said. “At the end of the day, where I believe all of this should be taking us is we need to be way more transparent … All the stakeholders in senior living need more transparency.”

By consistently capturing care data, Jambunathan said, the organization is “building a strong foundation for predictive analytics and more proactive communication with families about their loved ones’ well-being and daily care.”

Anthem measures the returns on AI investment through hours saved on a weekly basis by its team members and comparing it with the investment it makes for AI programs, such as the company attorney taking a document that would often take up to six hours to parse through or require an additional paralegal and is able to summarize it in a fraction of the time.

That level of time back is being seen across Anthem’s entire executive and community leadership team, according to Scott. If only one leader saved eight hours per week, it would justify the spend, he said.

The post Senior Living Operators Find Good and Bad in New AI Platforms appeared first on Senior Housing News.