AI Series - Part 7: Profile + Stage & Extended Technologies
I recently had a family member need extensive surgery so we went to three doctors in different cities to find the best options for the difficult journey. Ends up there are a lot of options and choosing the right one in the moment is tough. Your emotional and you don't know who is guiding you through the right path. We were lucky in our choice in the end but optionality is not really what you want in the case of life or death choices. Your healthcare decisions shouldn't feel like a gambling play in Vegas. Then the techie in me started to bounce the idea around. What value would AI add to address this challenge? Automation, collaboration (of data sets and best practices of the best minds and medicine globally!) and standards... "We can cut it out vs we can fix it" - big difference. So what does AI bring?
A consistent, streamlined, supported experience through the entire lifecycle
Improved customer engagement in the process
Greater customer Trust (Loyalty)
Improved company brand awareness (yes, medical center have brands).This is the Gold of AI. The first companies to figure this out and leverage this across their org will win and win big. The gun went off a long time ago - get busy people.
Profile + AI: The User Profile is the the BASE. Each User is defined by their own experience (data set), filtered so the company system can provide support via channels, 24/7 across regions and countries with built in rules, best practices and supporting information to best handle their top customer issues. And the big one here is that the more data that is collected, the better the customer is serviced. This is the definition of an ‘Antifragile System’, something I cover in depth in my articles and have made suggestions on reading (ie. Nassim Taleb - Antifragile: Things That Gain from Disorder (Incerto) - apply to technology).
Profile + Stage + AI: Now we are getting to the next level of data. (1) Who is it and (2) where are they in the process? Support each process and pass the baton to the next experience which contains new people, a new center, new rules, new mindsets etc. ie. Surgery and Post Surgery are handled by completely different hospitals, systems, people and the experience could not be more different with one hospital behind on key medicines and diagnosis options. Why?
With AI, when the patient begins to engage the medical center's AI platforms (ie. Chatbots, KB, etc) there is relevant content and answers based on not just their past engagement and profile but their present stage and other patients experiences. This adds another layer of key data to push relevant information to support the patient in their journey.
Clarity is what you want. Focus on the Stage. This is where you can take your metrics to a new level and apply across different channels. Again, each stage may need a new channel. Explore, break, apply, score!
AI Model Overview: Now before we go further let's review what AI models are: An AI model (*n) is a program that analyzes datasets to find patterns and make predictions. AI modeling (*v) is the development and implementation of the AI model. AI modeling replicates human intelligence and is most effective when it receives multiple data points (*like a human) per the ‘stage’ of medical treatment the client is in as stated in the example above.
Implementing AI Steps: Here are the steps to implementing AI models. We aren’t quite there yet but its good to have this set in the back of your mind as we go through the next section and I start making references to algorithms, platform choices and such.
Choose the 3-5 Use Cases (from the 10) and map correct requirements (discussed in last article)
Choose the Model(s)
Create algorithms
Train the AI model
Choose the right platform, infrastructure changes, apps
Pick a programming language
Deploy
Monitor the operation of your AI system.
Told you it wasn't easy. Can I make a suggestion...go back and create some Predictive Models. Review and let me know your thoughts. It will help align your AI data.
Extended Technology: Remember those use cases back in Part 5 - B. By this stage you should have figured out the best base technology to use for your use case requirements. Now lets add another layer or two! Think about extended technology like IoT platforms (Thingworx), Remote Assistance (RA) (Techsee), Dynamic Checklists (Youreka), Project Management tools (Sitetracker), and other technologies that combined with AI provide an entirely elevated experience. Get out there and look for ways to add more 'supportive data' to the core process (use case) with extended technologies. There are some real winners out there.
Quick Actionable examples:
Automate routine tasks: Improves employee productivity to help reduce costs so time can be spent on more important, complex tasks. Trick here is you need to have identified the tasks, metrics of value, etc. Also, remember to apply those gray value areas like raised skill set levels of ALL employees. (I can give details if curious on why this is if its not clear.)
Automation of tasks such as basic inquiries. “How do I…X?” “What form do I use to …X?” “What do I need to do to…X?” Also, it will fill out forms, create actions, submit and update information. There are many options here.
Chatbots provide customers with quick and accurate answers to their questions. By answers I also mean ‘actions’ like setting up appointments, notifications, knowledge forms and articles, etc. Its endless.
Remote Assistance (RA). Think about RA (Remote Assistance). Customer: I have a problem with X. Chatbot: Let me get a technician on the line and provide some RA for a walkthrough of this particular issue.” Customer: Sweet! I love AI.
Voice Recognition tools can automate tasks like scheduling appointments and booking services.
Predict customer behavior with access to history and pattern recognition. Like amazon but more advanced in its algorithms, if you like this +you are this demographic + you have bought this product we will provide options for X and based on profile, history, you are in this STAGE, we will price it out accordingly. I only use this example as I just bought an exercise product and I’m still getting advertisements 3 weeks later for the same product I bought. Really? What about providing me one of your partners for supplements for my age which I told you, or a group in my area (that you know) to join that uses your products in their gym. So Lame! We are still thinking about what's in it for me, me, me and its limiting sales tremendously.
Fraud prevention is a great use case that I want to mention here. AI does this job more efficiently than humans like detection of workers' compensation fraud. Outcomes lower loss costs and total risk costs. There are serious numbers in this area. I'm going to go into some great use cases later where AI is KING, or QUEEN and humans will just say - you Won AI - take it from here.
A designer note on ChatBots: ChatBots are amazing when done right. Then again, I have been online and had some not so successful bots to work with that provided so much text and context that it was so overwhelming I forgot the reason I reached out. Ask a few chosen clients, customers and employees so that you don’t impact the value of your message. No one will complain, they just won’t use that channel and you won’t know why. Ouch. Your teams will think AI doesn’t work when its really just the context and strategy of the message. Validate, Validate, Validate all the way through.
Well congratulations if you’ve made it to part 7. Thank you for reading and hopefully this has been a value for time spent. Ian Moore