by Chris Pehura — C-SUITE DATA — 2024/11/17
Reach out if you want to learn more
When you talk about adopting AI in your company, how does that conversation go?
Well, it can go in several different ways. You could talk about the benefits of AI without getting into the nitty-gritty of how to actually do it. There’s some risks there. Or you could talk about the AI practitioner and how someone would use AI at your company. Or you could talk about specific industry applications and how AI is being used in other industries, including your own, and how you could bring it within your company, those type of applications.
No matter how the conversation goes, there’s always a sort of a hole in terms of these conversations. A lot of people don’t have a common understanding of AI or a common vocabulary for it. So when you have these conversations, they can be relatively bumpy. So when you want to adopt AI, you got to first focus on building a common understanding and a common vocabulary so that you can move the conversation forward. Then you can talk about investment. You can talk about redesign of processes. You can talk about training. You can talk about even hiring the right people because you have that common language, that common vocabulary that is not technical jargon. Technical jargon is too ambiguous, too nebulous.
And so you talk about the patterns. You talk about the patterns for AI adoption. You talk about how people in different industries are following these same patterns and how they’re adopting AI. And once we have some patterns, some core patterns, we can use those patterns in combination to develop more elaborate patterns, more elaborate vocabulary, and how we can better understand and use AI within our companies. So let’s talk about some of these patterns.
Approximation pattern. This pattern is where AI is used to develop a solution, not the right one, not the correct one, but a rough solution that should be reworked.
Optimization pattern. The optimization pattern is where AI is gathering data, then it reorganizes, re-categorizes, refactors that data so it better understands and uses that data.
Execution pattern. This is where AI is used to conduct an assessment and then AI may recommend or take a next step.
These are the three core patterns. And with these three core, we can create all these combinations of more elaborate patterns when we’re discussing how AI is adopted in our companies. So let’s get into a more deeper into these patterns.
approximation pattern
This is where we are getting examples. We’re putting the examples into a template, and then we’re fine-tuning the template, tweaking it.
Now, when people are doing individual, unique type of work, this is how people normally work. If I wanted to put together a strategy, I would look at examples in industry, I would then put together a boilerplate, and then I would then vet it with my team. If I’m designing a logo, I’d be looking at what the other companies are doing, getting examples, I’d start a draft, which would be my template for my logo, and then I’d vet it with my team.
This AI would support people like that, the approximation pattern would support people doing those activities. Now let’s get into the next pattern, the optimization pattern.
optimization pattern
The optimization pattern is where data is being gathered, and then data is being reorganized, recategorized, factored, and then the data is adjusted based on what was learned.
People work like this. Analysts work like this. People who do data entry work like this, where they’re looking at data in free text and then they’re entering it into a form.
I am in customer service, I get customer complaints. Well, what do I do? I see all this text and then I’m entering in my report and what the customer’s really complaining about. And maybe do an analysis of what really are the top issues or top complaints that we have for customers in a specific segment. I’m doing this manually, and then based on what I’ve learned, I am adjusting the data so that it makes a lot more sense on what’s going on.
What some companies do is that they want to submit their invoices during the supply chain. So rather than hard wiring their data to specific tables and table structure, they have free form of their invoice and then they just submit the invoice to a company. Then the company has an AI where it’ll mine through all that information on the invoice and then automatically load it up into their into their systems, into their purchasing systems. You don’t need to code all those complex rules and mappings and worry about a million different combinations of invoices coming into your company. The AI handles it all. This is not just for invoices, claims, patient records, and even background checks. That’s the optimization pattern.
execution pattern
Let’s get into the execution pattern. This is where the AI does an assessment, then based on that assessment, it specifies a set of rules it needs to follow. It doesn’t go off right away off the assignment. It identifies a set of rules, the right rules, what it needs to consider and follow, and then it does something, a next step, which may be just notifying somebody or actually doing some form of automation.
This is how people who want to route work, route people, route mail, this is generally how they work, or even smart driving cars. That’s how they work too. To keep the car safe, like if you have a camera when you’re backing up and it shows you the camera. It’s doing an assessment of your surroundings and then might give you some rules in terms of a beep that you’re getting too close to something and some actions are need to be done.
I’m a patient in a hospital, I’m in an emergency, an assessment is done on me based on various criteria which may route me to various doctors, tests, or even a bed. I’m on the assembly line, a part’s coming down, the assessment would be looking at the part, its orientation, making sure there’s no defects, things like that where the AI may decide to reposition the part, have a robot do something to the part, stop the line, or let things move forward. That’s the execution pattern.
sum it up
These aren’t all the patterns. You could mishmash and combine a lot of these things together to form your own patterns. But the thing is that you identified patterns, you identified a vocabulary. Now you can identify things that you can actually talk about and scaffold to your existing processes.
When I look at these patterns, I have some ideas now on how to do process redesign, or how to do training, or who to hire, and when to invest. I gain a lot of value from using patterns like this. Now, these patterns can be misapplied, and that’s a very different conversation altogether. And they’re misapplied because people miss out on the nuances of these patterns, the nuances of how AI works, and the nuances of how people actually work within their company.
I’ll talk about all of that in my next video.
See more content at csuitedata.com/content
Want to chat live? Setup a time with Chris at csuitedata.com/contact-us