I was a guest on the Centric Consulting podcast discussing Artificial Intelligence with host John Kackley. John and I start basic but quickly dive into how AI is, and will continue, to be a disruptive force, for better and worse, in our personal and professional lives.
John is a talented facilitator, with a voice made for podcasting, so definitely take a listen.
John Kackley: (00:00)
Hi, this is John Kackley. I’m a project manager and program manager with Centric Consulting. And this is the second in our podcast series discussing trending topics and technology in business. In this podcast I’ll be talking with my colleague Kerman Fontana about artificial intelligence in business. Thanks for listening. So Carmen, tell me more about yourself.
Carmen Fontana: (00:18)
I’m Carmen Fontana. I am the Cloud and AI practice lead here at Centric. I live in Cleveland, Ohio. Outside of work with my very scarce free time, I love to travel with my husband and sons. I’m a runner. I like to read and try to do jigsaw puzzles too to keep the brains fresh.
John Kackley: (00:36)
You moved into this role and you’ve been working on a project with AI, at least, let’s understand. So tell me about it. How did you get involved with it? What’s it like?
Carmen Fontana: (01:41)
Yeah, so we started out as a proof of concept, doing it, practicing what we preached and did it on ourselves and saw that there was certainly some legs to it. But we wanted to apply the concepts that we learned from doing it internally on some different datasets, organizations that have much higher turnover rates than we do at Centric. We’re kind of blessed here that we have pretty low turnover rates. But, some of our clients have major pain points, you know, 20, 25% turnover. So we have two different clients we’re working with right now, both work in industries where they need skilled workers, but there’s a lot of competition for those skilled workers. So we’re hoping that this project allows them to figure out, the best type of people to hire, but also how to retain those top hires.
Carmen Fontana: (00:43)
So I’m working on a project we’re calling it the employee attrition project. And the premise of this project is really how can we leverage machine learning to analyze employee patterns of behavior and ultimately predict which employees are most likely to leave your organization. And really, you know, the true motivator for this project was, you know, we were seeing it both in ourselves but with also with our clients that the unemployment rates are painfully low. Painfully low if you’re an employer. Probably good if you’re a candidate on the market. But it’s really hard right now to attract and retain the best talent just because there’s so much competition. And so being able to have a smart, intelligent way to identify potential hotspots in your organization, allowing you to be more proactive about some of your human resources practices. It goes a long way and it’s a business driver and in this job market.
John Kackley: (01:34)
Okay, well that does sound exciting. So was this something, are we doing this for a client or is this something that we’re doing internal to Centric right now?
John Kackley: (02:26)
All right. And just to go a little bit deeper in it, cause you’ve got me intrigued, how is AI different in a case like this than sort of typical data analytics? Because it sounds like a data problem, right? But how does AI, how does AI make a difference?
Carmen Fontana: (02:44)
Yeah. And you know, that’s one of the questions we get is managers say “I know which one of my employees are going to leave”. And I always say you probably do because most good managers, they know their employees, they’re talking to their employees. They have a good sense of which employees most likely to leave the organization, but that doesn’t scale real well. And so that’s where machine learning is really helpful. You know the six people on your team, but you don’t know the 6,000 people within your organization, their propensity to leave. The difference between predictive modeling, machine learning versus traditional analytics is traditional analytics tends to be retrospective. So looking back at what happened in the past, and certainly there’s an element of that with machine learning, you need to see what happened in the past. You can learn from it. But really machine learning is looking forward. So, “Hey, let’s learn from the past, but then let’s apply it to the future”. So let’s take our current employees and predict what they’re going to do in the future. I think that’s really a differentiator between traditional analytics and predictive modeling.
John Kackley: (03:44)
Okay, cool. That starts to put some shape around it. I like that. Uh, so let’s break it down a little bit more. When people talk about AI or machine learning in business or they’re talking about, you know, is it really one thing, you know, that we’re just seeing a lot of different applications. Is it three different things? Is it 10?
Carmen Fontana: (04:00)
A very common question because there is so much terminology and kind of all bleeds together and it’s complicated because there is so many terms. But honestly at the end of the day it’s pretty straightforward. Artificial intelligence or AI, it’s kind of become a really generic catchall term for just intelligent technology. So that intelligent technology might be image recognition, natural language processing. Common applications you might see are self driving cars or Siri or Alexa, you know, those are applications of artificial intelligence, intelligent technology. Another word which I’ve already dropped is machine learning and machine learning is kind of the engine that makes artificial intelligence run. It’s the brain behind it, so you’ll hear that word a lot of times machine learning and AI are used interchangeably. I try not to get fussy about terminology because at the end of day it’s about the intelligence. One other term you might hear a lot, it’s become very popular is deep learning. And I say deep learning is that machine learning with more sophisticated math
Carmen Fontana: (05:04)
I say that a bit facetiously. There is certainly more to it, lot more math involved, but, deep learning is really important development in this space because it allows us to do really complicated things that just really weren’t possible before. And when I say complicated things, you know, the image recognition and the natural language processing, that’s fueled by deep learning. So deep learning, machine learning, that’s the engine, artificial intelligence, that’s just a generic word.
John Kackley: (05:33)
So AI is really big right now. We’re seeing it discussed as a potential solution for a lot of different business problems. Where do you see it having the biggest potential and do you see any areas where you don’t see it playing much of a role?
Carmen Fontana: (05:44)
There’s a lot of excitement, a lot of case studies around how machine learning and AI is used in the consumer space. So, for instance, Netflix, they use machine learning to predict which movies you might want to watch. One of my favorite machine learning applications, StitchFix, it helps you predict which clothing that you might want to buy. There are a lot of really fun, exciting consumer application. But where I really see the biggest potential is in the kind of the non-glamorous stuff that businesses need to do every day to run better. One huge explosion area is using AI for predictive maintenance.
John Kackley: (06:22)
Carmen Fontana: (06:23)
Think solar panels out in the desert. Company spend a lot of money in the repair cycle. And being able to protect, predict head of time, what’s going to fail allows them to be smarter about fixing things but also ordering replacements, things of that nature. Especially not glamorous but a huge impact to the bottom line we’ve seen in our space, you know, once again talking about practice, what we preach. We use machine learning to help us staff our projects smarter and we saw a huge improvement in how we run our business. I think that’s where companies are realize the biggest value is when they apply it to some of the business processes that really drive how they work and go forward.
John Kackley: (07:03)
That’s a great summary there. But do you see places where you don’t see it playing much of a role or just doesn’t quite fit?
Carmen Fontana: (07:10)
Great question. People are always worried about “are the robots going to take all my jobs?” If the answer is they are going to take some of our jobs, there are jobs that are going to be eliminated or at least reduced by this technology. Whether it’s truck drivers because of autonomous vehicles or if it’s radiologists, doctors who read CT scans because the technology can do it as well as they do. There’s going to be impact in those areas. But the places where we as humans really exceed and I just don’t see the computers catching up in, are those jobs that requires creativity, empathy, and very nuanced communications. That’s where humans are going to outpace the robots, for a very long time. And so I think organizations, if they can use the AI to optimize business processes and then really lean into where they can provide value through creativity, empathy, communication, et cetera, those are the ones that are made the most successful.
John Kackley: (08:13)
Right? I can see a risk there, a company going a little too far in AI and, and automating things that take out the creativity and human nuance and, having to come back on that. So what advice would you give to an organization that’s asking themselves about using AI?
Carmen Fontana: (08:28)
Well, the first thing I would say is it’s coming. It’s coming hard and fast. So if you’re not thinking about AI right now, you need to get on that. AI is real and there’s been certainly hype cycles around AI in the past, This time it’s real, it’s happening. I would recommend if a company is new to AI, which many companies are totally fine. It’s pick an area of your business where you can focus, where you can do proof of concepts, where you can build out understanding of AI as well as confidence and comfort and familiarity with it. Use that and then figure out, “okay, where else can I use this in my business?” And let that snowball.
John Kackley: (09:08)
Where do you see the biggest challenge as an organization would face in trying to make use of it?
Carmen Fontana: (09:13)
Well, there’s certainly a lack of AI skilled professionals out there, so just being able to find the technical knowledge is, is difficult. In our case, a lot of our partners come to us to leverage us for those pieces, but we also recommend that companies have that AI knowledge both in house as well as the partners. And because it is going to become a fundamental piece of your business. I would the other piece, which is equally as important as the technology is really spend time thinking about the people. There is an amount of trepidation about artificial intelligence, and rightfully so, really. Your organization needs to brace for that change. It’s it’s not done well, it’ll be abrasive and it’s gonna be painful. But if it’s done well, they’ll still be nerves. But you can do it in a way that’s empathetic and knowledgeable about the humans that work with you.
John Kackley: (10:10)
Great. So you started touching on the technology. They’re a great lead into the next question. Are we still at the point of seeing package solutions for AI or machine learning or is it all pretty much homegrown stuff at this point?
Carmen Fontana: (10:23)
What we’re seeing right now, with the explosion of AI, is because AI is becoming way more accessible. So I actually got in this space, late nineties. My senior project in college was, in true college project fashion, very nichey, but, building a machine learning model to model climate change and in the former Soviet Union. So it was a bit ahead of its time. Unfortunately, we weren’t talking about climate change back then and certainly weren’t talking about Russia. Maybe we should have been. And would even use the word machine learning. But at that time, it was very difficult to do this predictive modeling. We didn’t have the compute power to do the complicated calculations that are required. It required a deep level of data science knowledge. Fast forward to today, we have the cloud which allows us to scale very quickly, and add ton of compute power so we can run our algorithms and be done with it. But we also have these pre-packaged solutions that do the natural language processing, that do the image recognition, for us. And so we don’t even have to be a data scientist to do that. AWS, Amazon web services, Microsoft, Google, they all have these built in cognitive services that can really be applied to your existing applications in a pretty painless manner.
John Kackley: (11:39)
Cool. What do any really stand out right now you talked about a different packages? Any that stand out as a leader or somebody to watch?
Carmen Fontana: (11:48)
Yeah. You know, AWS and Azure have really run out in front of the crowd in terms of their dominance with the cloud space. And their dominance of making AI accessible to the masses. We are seeing Google make a pretty big effort in that space. They’re not quite the same level of those two, but of course Google is Google. They have a lot of smart people there. They’ve been working in the AI space internally for a very long time. So once they start leveraging that externally for other people to use, I certainly see them as being a player as long with Microsoft and Amazon.
John Kackley: (12:24)
Okay. And relating to what it takes for an organization to deploy it. You started to mention a couple things and see a lot of of complimentary disciplines. Now, obviously big data is a necessary part of this, change management comes in. Do you see any other significant disciplines that need to be tied in and also need to be successful?
Carmen Fontana: (12:46)
Yeah. Well, you know, you nailed two of the most important ones. Data & analytics – these algorithms run on data and the more data you have, the smarter they are. So data and analytics is a natural partner with artificial intelligence. Change management – we already kind of talked about that. This is a big change. It is a change in how we work. So making sure we have an emphasis on the people is so important. We talked about the technology. We talked to the people. I’d also throw in the process. We use the example earlier about predicting equipment failures. We have the technology to predict the failures. But you still need have to have a process in place. Once you have that prediction, how are you going to surface that in a timely manner? How are you going to apply that to your supply chain and make sure you have the parts on time. So it certainly is an integrated model, AI, between people, process and technology.
John Kackley: (13:39)
And then you talked a little earlier about certain sorts of jobs will be displaced or certainly changed significantly, based on AI. Do you see any particular sorts of jobs that AI will create?
Carmen Fontana: (13:53)
Especially we’re already seeing tons and tons of technology jobs being created by AI and that’s just not computer scientists. We’re seeing a lot of opportunities for statisticians, mathematicians, people with that background. So technology is an obvious choice. Other interesting places where I think AI is going to create jobs is in the legal and ethics side. Right now probably our technology is a little bit ahead of our understanding of its impact. We have seen that in the news in many cases. If one of these self driving cars has an accident and somebody’s hurt, who’s liable? We don’t necessarily have that all figured out right now. That is where the legal piece is super important, as well as the ethics piece. It is one thing when a person does something to another person. But what if a robot does something to a person? What does that mean? I see that as an area that’s really going to expand over the coming years.
John Kackley: (14:51)
Random sort of question that comes to my mind from that… our country and our society generally drives technology pretty much as fast as far as we’ll go. That’s kind of to who we are. At some point do you wonder about what might trigger a backlash against AI? I don’t even know if you could corral it.
Carmen Fontana: (14:51)
I think you are spot on when you say, we embrace technology here in the United States and we want it to succeed, which great. There is some really great reasons why we want to do that, but sometimes we get out too far ahead of ourselves. Social media is a perfect example. That’s not a newer technology. It’s been around for five, six, seven years. But we’re just now understanding the impact of social media. There’s lots of talk about social media and our mental health and our physical health, but we’ve also seen that social media can be weaponized by bad foreign players. We’re just now figuring out what social media is good and bad for. So I worry with artificial intelligence. Right now we’re in the rah-rah stage where everything AI is good but maybe in five or six years we’re going to have that “Oh crap!” moment where we’re like, “why didn’t we think about this that back at its inception?”
John Kackley: (16:07)
It’s clear that you’re super excited to be working on AI and you make me want to go join your team. Do you have any openings? What about an AI excites you the most?
Carmen Fontana: (16:26)
We’re always hiring. Send me your resume! I think for me it’s the potential to solve problems that are going to help humanity. We’ve certainly focused on the business stuff today, but at my core, that’s where I want AI is to help the humans. Personally, I’m cancer survivor and having been through that experience, I know that the treatments are pretty generic. They’re getting better, but they’re certainly not tailored to your age, your medical history, this or that. They’re doing some really interesting things with AI to really create customized treatments down to the specific patient based on all those different factors to lead to the most positive outcomes. This stuff really excites me because it’s at its core, it has a very pure purpose because it is trying to make us better as humans.