By Jack Cumming
The explosive arrival of artificial intelligence (AI) into public awareness has raised both fears and hopes. The skeptics have questioned the ethics. While closer examination makes the dark fantasy of “2001: A Space Odyssey” seem unlikely, there are ethical questions which we’ll take up in a subsequent article. Here we consider whether the hopes are well-placed.
Suffice it to say that existing ethics and legal processes can readily be applied to AI, and they should be. This article focuses on AI and not on its application in robotics. We alluded to that in an earlier article. In this article, though, we consider what AI does well, what needs more work, and what may never be achievable. These are practicalities.
AI is good at handling routine, recurring, mechanical tasks, and it’s also good at providing responses etc. that mimic humans even though the responses are mechanical. This can be particularly true for those myriad situations in which statistical tracking and prediction can improve efficiency and save money.
A classic for senior living would be menu planning and ordering to provide resident favorites while staying within a budget and minimizing waste. Other applications leap instantly to mind, for instance, collecting and analyzing data to minimize maintenance staffing and expense while maximizing resident satisfaction.
Best to Avoid
Because OpenAI ChatGPT initially seems uncannily conversational, it can be tempting to substitute AI solutions for interactions where customers expect human contact. It’s not long, though, before its stiltedness comes to the fore. The ability to convincingly mimic real people may come as AI development continues. For now, though, AI is still rather stiff and rules bound and so is anything but the warm reassuring presence of an actual human. Thus, if you are trying to entice customers, it’s best to avoid using AI unless you disclose that it’s not person, and you make access to a real person available at a finger touch. Click here for a deep dive into AI capabilities.
Another thing to avoid is any vendor who touts AI as a reason to consider their offering without spelling out how it will solve your problem. If you don’t understand the offering, and how it will fit into your organization, and why it is better than alternative vendor offerings, don’t waste your time schmoozing with a sales rep. What’s more, if you can’t interact directly with the development engineering staff, don’t pursue the contact further.
As a buyer, if you don’t have the math and other skills to understand the engineers’ explanation, then find someone who is. This is not a purchase that is well adapted to the feel-good school of salesmanship. Unless the salesman is also a developer, the sales process is just a time sink and overhead cost and is best avoided. AI solutions should be bought by you and not sold to you.
Buy or Build?
Still, there are many untapped areas where AI has the ability to make a real difference. Those applications may be developed by vendors, but there is also the possibility that a senior living provider organization with sufficient skill can better develop its own application. After all, providers are in the best position to know what the opportunities are and how best to tailor a system to those possibilities.
If there is insufficient internal staff, a provider can contract with gifted individuals. Contracting requires a provider executive with the insight and ability to identify the kind of genius that can give a satisfactory result. Not everyone in a managerial role has that kind of ability to recognize the best talent. Some talented developers are edgy or introverted. Getting that match exactly right is a key challenge for senior leadership.
My personal favorite example of how a leader ensures effective talent is how President Roosevelt chose Leslie Groves to manage the Manhattan Project. Groves had built the Pentagon, and Roosevelt was impressed, so when he needed a no-nonsense person to manage the scientists, he turned to Groves, who was a lowly Colonel. Roosevelt had to promote him to General on the spot to give Groves the gravitas to get the job done. Organizational success depends on CEOs with a similar eye for talent and with the drive to act decisively on that recognition.
The Shy Prospect
Here’s an example of what AI now can do well. Consider an initial website contact from a tepid prospect who is just dipping a toe into a relationship with your community as a home for his or her folks or him- or herself. That person is interested, or they wouldn’t have visited your website in the first place. You want to keep them interested and build a relationship. Today’s practice of immediately asking for the prospect’s name and contact info drives many people away.
The last thing in the world that a shy consumer wants from your website is a conversation with a live salesperson. He or she just wants information. He or she wants to get a sense of whether your organization is trustworthy. And, he or she wants to be able to think without sales pressure. It’s a much more important choice for the prospect than for a salesperson.
That’s where AI conversation can be a big help. Consider North Carolina where there is a tremendous amount of data in an extensive disclosure document which is available publicly for every CCRC in the state. If we let our hesitant customer know that he or she is only dealing with AI, i.e., a machine, he or she is likely to stay longer on the website than if we immediately ask for his or her phone info so a salesperson can follow up.
Training AI on that mountain of disclosure data can give our shy prospect mountains of information to keep him or her in discovery heaven until he or she has sold him- or herself. Then, when he or she finally does allow a conversation with a live salesperson, he or she is ready to make the buying decision. That means that his or her fear is already assuaged that he or she may be duped by clever sales techniques into something regrettable.
The lengthy conversation with the AI-persona helps the prospect to get to know the organization and to learn whether it is to be trusted or not. Transparency, like that required in North Carolina, builds trust. Trust is the main element that is essential to growing occupancy.
Just as the internet revolutionized everything and fostered more connectivity and knowledge, so AI has the potential to eliminate many mindless, repetitive tasks, freeing humans to have more free time and fewer mind-numbing, repeating chores. Like both the internet and humans, though, AI is capable of both ethical and unethical uses. We’ll take that up in a companion article.