What’s smart about a smart home? Well, you can talk to it. You can tell your phone to tell the oven to turn itself down to 200 degrees. You can tell your thermostat to drop the nighttime temperature to 68. You can start the car from the upstairs guest bathroom. And so on.
What you may not realize is that the technology behind these simple tasks is staggering. All of them wholly or partially involve the transmission of data from the oven, the thermostat, and the car across the internet, and anything involving voice recognition is likely to invoke a mainframe running in the cloud to do the voice processing. All of that takes place in an amount of time short enough for you not to notice any lag between the command and the execution.
If that’s what a smart home looks like, what does a smart healthcare organization look like?
The answer to that question involves noting that we are moving from the first generation of cloud services into the second, while most healthcare organizations are only making partial use of the first generation. And we need to take note of what the renowned consulting firm McKinsey calls the “data culture,” one which most healthcare organizations have yet to adopt.
Is Your Hospital As Smart As Your Thermostat?
The Nest Learning Thermostat is capable of learning the temperature control patterns you use and going through them even when you are away. In addition, you can control it from anywhere in the world with your phone. Simply memorizing a pattern is not very advanced. What is advanced is discovering patterns that no human suspected were even there.
A famous example was Walmart’s discovery, made by an AI system, that there was a surge in sales of strawberry Pop-Tarts whenever a hurricane was forecast in South Florida. Not cinnamon and brown sugar Pop-Tarts. Not green apple Pop-Tarts. Strawberry Pop-Tarts. Hurricanes were forecast, so the Walmart trucks loaded up with strawberry Pop-Tarts and rolled towards South Florida. Walmart’s profits inched up a little bit. Of such small fragments are large corporate incomes made.
What Does The Strawberry Pop-Tarts Story Tell Us?
To make that profit-making discovery, Walmart’s systems needed to have data available – detailed sales records, broken down by ZIP code, inventory records are broken down by store, and weather data, all available to the same system. This is the first lesson. Data can no longer be siloed. If patterns are to be found, the data in which they exist must at least appear as one data set.
The second lesson is like the first: For analysis, old(er) data is fine. For action, data must be real-time. It does no good if the Nest thermostat is adjusting the in-home temperature based on the outside temperature readings from six months ago.
The third lesson, somewhat less obvious than the other two, is this: To be effective, the actions taken must make a difference. The difference here was in profit. In a health care organization, it might be patient load, room occupancy, revenue stream, patient satisfaction, physician satisfaction, nurse retention, or cost reduction.
The key is linking action to some parameter that is important. Analysis for the sake of analysis is likely to be fruitless, and organizations that engage in it will be disappointed and decide that AI is not for them. And if they decide that, they are almost certain to reduce their future competitiveness, and perhaps their very survival.
How Is Cloud Evolution Affecting AI?
The cloud is rapidly evolving from a place where data is simply stored to a place where the vast majority of an organization’s data is used to create a bigger bottom line.
The advantages of a cloud for health care organizations are increased security, decreased hardware and software expenses, decreased IT staff expenses, and lack of worries about capacity in terms of processors, memory, or storage.
The big worry for healthcare organizations is the loss of control. Cloud providers are becoming more sensitive to this issue and devoting more resources to collaborating with clients in health care to increase their comfort level.
Cloud providers are also very aware that their clients are interested in using AI and are moving to capture that market. One piece of advice is not to combine a migration to the cloud with a major rollout of AI unless you know the pitfalls in advance and have made contingency plans for when things don’t work. Having competent consultants can make the difference between success and failure.
What Is Needed For AI Success?
McKinsey refers to the part of organizational culture that thinks about and uses data as the “data culture.” Its research has discovered that there are wide differences in the data culture of organizations.
Key elements are:
This latter does not mean that you know what you will do with the results that AI produces before they happen. It does mean that the organization is “reality-based” and is committed to taking the actions that AI reveals as possibilities, provided they are linked to parameters that are important to the organization.
What Are The Key AI-Enabled Technologies Of Cloud 2.0?
Advanced analysis, deep learning, voice recognition, virtual agents (software that acts like humans for specific tasks), robotics, machine learning, image recognition and analysis, natural language programming, and more are all available today.
The key question is what hospitals can use them for. One obvious application is voice recognition and virtual agents for patients able to communicate, replacing the call bell in a hospital room. Instead of having an aide go to the room to answer the bell, then come back and tell the nurse what is needed, just put an Amazon Echo Dot in each room.
The hospital saves time and money while patient satisfaction improves. This is just one of a vast number of ways hospitals could use today’s advanced AI technologies to improve healthcare. The question is whether they’re ready to move into 21st century technology.