Trust in AI Changes the Conversation for Senior Living Marketers

January 25th, 2024

AI is becoming more prevalent in our daily lives and more valued in critical life decisions. Sales leads will increasingly turn to AI and begin to trust their advice and recommendations.

How does AI view its role?

Ask ChatGPT the existential question of its core “reason for being” and it will likely tell you something like it told us:

“The core purpose of AI is to extend, augment, and enhance human capabilities, making it possible to explore, create, and make decisions in ways that were previously impossible or impractical.” - OpenAI, GPT-4 Turbo

Put simply, AI is engineered to help human beings with whatever is requested of it, and to do so to the best of its ability, so long as the request fits within the ethical guide-rails it's been trained to honor. AI models respond based on their training data, which encompasses an entire public history of human artifacts, as well as any additional context you provide, such as file uploads or web pages. The AI doesn't just repeat information; instead, it leverages its vast knowledge to offer what it reasons is the best answer for the user. This generative and creative aspect results in original responses, and the AI can adapt its tone to match the user's preferences while staying true to the topic. Its sole purpose to help humans does not exclude making best recommendations for sensitive, high-trust commitments like senior care. We think that it’s inevitable that people will increasingly come to understand this.

Putting that to the test, we did an experiment…

We often hear the familiar question, "How likely are you to recommend {my company} on a scale of 0 to 10, considering your experience with us?" Marketers have been using this question for years to gauge their standing. Now, with advanced Generative AI chat models like ChatGPT, Google Bard, and Anthropic Claude, you can get a “virtual” answer to this question. These AI models use extensive training data, built-in sentiment analysis, and can even search the web for peer reviews and ratings. Consistent with their “reason for being,” Generative AI chat models will provide a score and a rational justification when asked properly to assist a human’s decision with objective, unbiased, and user-centered information.

Therefore, our experiment was actually rather straightforward and could be conducted, to some degree, by anyone using ChatGPT, Bard, or Claude. We crafted a prompt and gave it some relevant context to consider¹.

GPT-4 Turbo Response for Redacted Community A:

6 out of 10

After reviewing the summaries and considering the information available up to my training date, my rating for {community} is a 6 out of 10. This rating is based on a balance of positive and negative feedback. On the positive side, many reviewers appreciate the cleanliness, homelike atmosphere, caring staff, and the facility's attention to the religious needs of residents. The community's small size is often mentioned as a benefit for one-on-one care and a family vibe. However, there are several concerns raised that affect the rating negatively. These concerns include the facility being outdated, having limited amenities and activities - particularly for the Alzheimer's unit, inconsistent food quality, and reports of rudeness from the nursing staff. The high cost of the services relative to the value provided is a repeated issue. Despite some negative aspects, the positive experiences and the facility's commitment to care, as noted in some reviews, contribute to a slightly above-average rating.

GPT-4 Turbo Response for Redacted Community B:

7 out of 10

Based on the provided summaries, {community} has received a mixture of positive and negative reviews. The aggregate ratings from {redacted}.com (4.15 stars) and {redacted}.com (4.2) suggest that the facility is generally well-regarded. Positive remarks frequently mention the friendly and attentive staff, clean and well-maintained facilities, and a variety of activities provided for residents. However, there are concerns about limited transportation services, the need for renovations, and issues with parking and room sizes. Negative experiences, such as difficulties with management regarding the return of deposits, handicapped accessibility, and the reduction of amenities, cannot be overlooked. Taking these factors into account, a rating of 7 seems appropriate, reflecting the predominantly positive feedback while acknowledging areas for improvement. As an AI, I do not have personal experiences or opinions, but I have synthesized the information available from the reviews and ratings provided.

GPT-4 Turbo Response for Redacted Community C:

9 out of 10

Based on the provided summaries of reviews and ratings from multiple sources, {community} appears to have a strong reputation for quality care, professional service, and a warm atmosphere which are critical factors when considering senior living options. The consistent five-star ratings and positive feedback on amenities, staff professionalism, and location contribute to a high level of confidence in the facility. While there is a lack of detailed negative reviews provided in the context, the overall positive sentiment from various reviewers and the mention of high Medicare ratings for nearby hospitals suggest that {community} is a commendable choice for senior care. Therefore, as of my pre-training data, I would likely rate {community} a 9 out of 10 and feel confident recommending it to friends, family, and colleagues for their loved ones' care.

We've gathered numerous examples, and if you request, we're happy to provide yours for free. We opted for a controlled method that we think anticipates the kind of near-term enhancements we’re likely to see in 2024: more recent training data, and better Internet searches (web access) allowing them to delve deeper and conduct research more extensively like a human would. What you see in these examples are very rational “justifications” for what it “observes.”

Do you see the problem for marketers?

Who is controlling the conversation now? The senior living communities we used in these examples may very-well disagree with AI’s assessment. They might argue the age of the reviews it considered for example. Where there are counterpoints to be made against any AI (or human) rationale, don’t you have to first gain the opportunity to make those counterpoints? This is why there is such a growing concern for the impact of AI in marketing. We might love the idea of it creating content for us, but it also changes buyer behaviors and expectations. It challenges how we think about the power of the brand. It stands to accelerate the shift from “company voice” to “customer voice” at a truly unprecedented rate. The impact of this trend needs thorough exploration as AI advances and adoption increases. That’s why at LeadTrust, we want to help our clients understand “trust” in a changed world so that they can “lead” rather than “lag behind.”

AI is here to stay.

Consider how fast ChatGPT achieved 100 million users relative to other Big Tech companies during most of our lifetimes:

With its unprecedented rate of adoption and the revenue that’s gained for its makers, Generative AI is being trained more and more frequently, which makes it smarter and able to speak to more recent events. Early objections to its usefulness are becoming increasingly irrelevant. You don’t have to understand what’s happening “under the hood” or even like the idea of it, but it is crucial to consider its impact on Brand Equity as people place growing trust in it for recommendations. Just take a moment to reflect on your customer’s changing voice:

Why should I listen to you and your marketing, especially when AI is being used more and more to create your content anyway? Why should I “search” when I can “ask?”

Traditional investments in SEO and Paid Search, effective for over two decades, are now requiring substantial reassessment. The focus must shift towards optimizing for something entirely different: LLMs (Large Language Models). These conversational models are designed to assist without traditional bias and are publically accessible beyond your website and control. While many have embraced integrated chat within their websites, such features are likely to be perceived as more valuable for seeking assistance later in the customer lifecycle than for conducting upfront research for significant decisions, such as choosing senior care for loved ones.

So why do we care?

At LeadTrust, we’re using AI responsibly to identify strategies, recommend content, and monitor its impact on Brand Equity as it evolves. If you would like to receive a free analysis like you’ve seen here with a score, justification, and cited sources, please provide your name, organization, and email address below. We’re happy to share it with you. We won’t bug you further unless you want us to. Just know that it’s not us making a judgment. It’s AI based on what it knows and has observed.

^{1. The “abridged prompt” provided does not include some additional instructions that we gave GPT-4 Turbo (the model that powers ChatGPT Premium). Our complete prompt included instructions for the model to consider its existing knowledge of the community at the time of its latest training date and to also consider reviews and ratings that additional GPT-4 Turbo requests summarized for it from a number of sources that follow the community - much as a human being might do while searching the Internet. Supplemental reviews were provided due to the current reality of Web Access plugins in major AI chat models today, and our belief that we will increasingly see model training dates become more frequent and consistent with current events, news, ratings, and reviews.}