VP of Corporate Development Glen Hilford joins Access Answers to discuss “AI for Business” – the hype around AI, organizational readiness for AI, and his upcoming webinar.

Also available to listen on:

PODCAST

TRANSCRIPT

Episode 21: Demystifying AI with Glen Hilford

Julia:

Welcome to another episode of Access Answers. I’m your host, Julia Vergara with Angela O’Pry. And today we are excited to discuss AI with an AI expert.

Angela:

None other than Glen Hilford, our Vice President of Corporate Development. As Julia mentioned, he’s an AI expert and also a wine connoisseur. So we’ll make sure to get some recommendations from Glen at some point in the podcast. Welcome Glen.

Glen:

Thank you, Angela and Julia. It’s a pleasure to be here today.

Angela:

All right, Glen. So before we get into all exciting things AI-related, tell us a little bit about who you are, how you came to be an AI expert and what led you here today?

Glen:

Well, it’s been an interesting journey. I started in the technology field and primarily worked in industry before joining Access Sciences. While in industry, I was exposed to AI, even though at the time I didn’t realize it was artificial intelligence. It was just a unique and somewhat novel way to find value out of information. And we’ve done some amazing things with it in industry. After leaving industry, I joined Access Sciences and brought that information or that knowledge into the organization.

Julia:

So today it’s clear that AI is everywhere. You hear about it constantly at home, in the news, in business. What are your thoughts on its popularity or hype as it will relates to organizations and business?

Glen:

Technology hype is not unique to AI. We’ve all experienced this. As we’ve looked at technologies in an attempt to address business problems and increase business value, the value of the data within business. Unfortunately, there are a number of software product companies who are using AI to hype their products. In some cases it’s valid. In some cases, they have either developed an AI solution that brings value to organizations or have enhanced an existing product doing the same thing.

There are other vendors who will slap AI onto a marketing message and make claims that are somewhat questionable. So one response to your question is the hype may be real or the hype may be completely false. And it makes sense to do a deeper dive into the actual requirements for what you’re trying to achieve and what the product has the real capability to do. We call this shiny object syndrome.

Julia:

Perfect way to describe it.

Angela:

I know you’re not dogging marketers over there. You yourself were over marketing at one point.

Glen:

I think I’ve lost that on my resume.

Angela:

So what you’re talking about, if it’s appropriate or not appropriate or really is the solution, kind of leads us to our next topic about viability. And viability assessments is one way that a company could really take a step back and say, does this really make sense for my business or not?

Glen:

That’s an interesting topic, and it’s something that’s at the forefront of my mind. I’ve spoken to a number of people over the last year or so who have come to us and asked about how to get started in the AI world. And one caution I would throw out there is that not all problems, not all business problems, are solvable using AI. Now I am an enthusiast. I really like the technology and have seen great value come out of its implementation. But there are just some business problems where it doesn’t make sense. The challenge for businesses is it’s easy to overinvest in AI without really understanding if the problem is solvable, if the problem is viable using AI.

The analogy I like to use is when you’re doing an acquisition or divestiture, but primarily an acquisition, and you’re doing due diligence before you’re investing money in buying a company, investing a little bit upfront to understand if the problem is solvable, if the requirements have been defined properly. And if the technology is available to address those requirements, if you’ve got that match, it’s wonderful and that’s money well spent and that’s money you would spend anyway. The other side of that coin is moving down the road and investing a million or two or three, and then realizing that the problem really wasn’t solvable to begin with, and that’s a tragedy. So that’s something we’re trying to help avoid.

Angela:

So even if you go through the viability assessment process, and the end result is that this AI technology does in fact make sense for the business problem you’re trying to solve. Does that mean then that the organization is ready to implement that or adopt that? Is there still an element there of your company may not be ready?

Glen:

Oh, absolutely. This is another big challenge that businesses are facing today. AI has the potential to change people’s jobs for the better. But change is hard and getting people within an organization to recognize the value of AI, especially when it’s automating part of a job they are familiar with, is a challenge. The premise and the promise of AI is that you’re going to automate rote mundane tasks and free up people’s time to focus on more valuable activities. But getting that concept across and getting people to accept it and embrace it is a real challenge within organizations.

Julia:

So, can business leaders expect more pushback when it comes to change management for AI versus change management for a non-AI technology?

Glen:

I think that’s a fair assumption. The other challenge with change management in the AI realm is that it is somewhat different. Here you’re taking work that people are familiar with and you’re taking that away from them. You’re asking them to work alongside a machine if you will, or a computer, and trust the results of the computer’s actions. And that leap of faith, the trust that we’re talking about out, is difficult for some people to embrace.

Angela:

Yeah. I’ve often told Julia, even in marketing, we need to be Prosci-certified. Change is difficult, no matter what function in the organization, you said.

Julia:

Well, I can understand where the pushback comes from, because Todd, a while back, sent some AI software that can apparently write blogs. I was like, “Why are you sending me this? You trying to replace me? You trying to get rid of me?

Glen:

I wouldn’t worry too much about that.

Julia:

We tested it out. It doesn’t do the best job. So I feel pretty safe, at least for now.

Glen:

Okay.

Angela:

Yeah.

Angela:

So Glen, when we talk about AI, there are a lot of things that are way over my head. If someone in a business decides hey, this is viable, our company’s for it, what level of expertise do they really need in order to try to implement this as a solution?

Glen:

Are our business analysts AI savvy? What I mean by that is business analysts can be very talented, but if they don’t understand the capabilities of the AI space, then they don’t know how to analyze a business problem, such that it can be addressed using AI or not. This goes back to the viability question. Do these resources have the skillset, the ability to discern what is viable and what is not, and to describe a business problem in a way that we can make that determination. So they have to have some level of AI knowledge, even if they’re not experts. That’s the first bullet. The second, there are some hard skills that are required. Data science skills and AI domain expertise, especially when you’re talking about machine learning, those tend to blend. It’s a union, if we’re thinking about a Venn diagram, and those skill sets really are necessary.

There’s a lot of marketing hype about having business people build these things directly. And that may be a reality 10 years from now, but today it’s not. Can you develop that expertise or can you organize that expertise in the context of an AI program? And the term AI program is not something that most people in the AI business are familiar with. We take that from the information governance world. But if you think of an AI program as something that surrounds and supports AI projects, then it makes more sense. How do you govern the projects and the products that they generate? The initiatives, the outcomes, the AI applications, how do you curate the data in a way that it’s protected? How do you ensure that data that feeds these AI programs is available, it’s accurate, it’s timely.

How do you govern the combination of data, the AI model itself, the application, and the results set in a way that makes it a unit, so that you can explain how it came to this outcome, explainability. And then you wrap change management into that. How do you support these initiatives moving forward? How do you communicate what’s going on? How do you coordinate the efforts of the business analysts and the AI architects and data scientists to the best effect for the organization? Those are all things that require expertise and knowledge and some structure so that you have a cohesive program that you can move forward. Otherwise, you’re just firing off projects with little or no coordination and no way to ensure that they are sustainable over time. AI is not one thing. It’s not monolithic. It can do things like predict the future or classify information or discover patterns of information within a batch of data.

It can automate processes intelligently. We’ve all seen chat bots, those are kind of simplistic. It can extract information out of documents in a cohesive way. It can recognize visual objects. We’ve all seen facial recognition on our cell phones. So when you ask that question, what level of expertise is required? It really depends on what you’re trying to achieve. For many of those use cases that I described, especially the ones focused on machine learning and object recognition, there’s quite a bit of expertise that’s required. You need some data science, essentially the people that look at data and determine, and it’s more of an art than a science, what data is required to solve a business problem.

You need AI architects and engineers in order to configure the software, train it. And equally importantly, this goes back to the program discussion, sustain it over time because the world is not static. The world changes. The business world changes, especially, and as conditions change, these AI applications or solutions need to evolve with those changes. So these people are all important. We’ve already talked about the need for trained business analysts who have not only a business background, but also some acumen and knowledge about the AI domain so that they know how these solutions might be applied to a business problem.

And of course, change management. Change management is always a huge issue with AI and being able to insert or inject that discipline as you go through a project early and often is very important to creating effective change within the workforce.

Angela:

Glen, you and I have had many conversations, water cooler-esque conversations about AI and something that may be happening is overthinking AI solutions. There are applications that could be simple, but yet valuable to organizations.

Glen:

Angela, that’s an excellent point. When I think of AI, and I think this is common with many people, I think about flashy technological advances, things like automated cars. But in reality, many of the business problems, we call them use cases, that bring enormous value to businesses are fairly mundane. Things like determining what the demand for natural gas is going to be in Chicago two days from now.

I can’t think of anything more boring, but at the same time, there’s enormous value in that to the organization that’s trying to plan how to get that gas to market. So going back to the shiny object discussion we had earlier, if we look at business problems, we should look at value and really avoid focusing on how flashy or how shiny they are. In many cases, mundane business problems, if you can solve them, can bring enormous value to an organization.

Angela:

Okay, Glen, we’re not letting you off the hook without a wine recommendation. What you got? Give us your pizza wine and explain what that is.

Glen:

Okay. I’m married to a wine snob. There’s just no other way around it. And we have come to an agreement that we have two classes of wine. One we call pizza wine, which is roughly $20 a bottle or less, so that Glen doesn’t feel like we’re breaking the bank to open a bottle. Our recommendation this week is a Pinot Noir out of Sonoma County in California. It’s called, I think it’s called Tuli, T-U-L-I. We’ve had several bottles of it over the last few months and have really enjoyed it. And I don’t feel guilty about drinking it.

A slight step up is a Cab, Cabernet Sauvignon, out of Napa, which is just one valley over from Sonoma if you’re familiar with Northern California. The name of it is Textbook, and it runs about $30 a bottle. And we have enjoyed that as well. I feel a little guiltier about drinking that, but that’s okay.

Angela:

Well, I like both of those and we’ll definitely have a happy hour after your webinar in a few weeks.

Glen:

Well, that’s great. I’m looking forward to that.

Julia:

And for our audience members out there, you’re going to have to invite them to the big event, tell us what’s happening.

Glen:

Well, you’re right, Julia. We are going to do a deeper dive into AI for business, specifically looking at demystifying AI, making AI approachable and understandable for business people. Our first episode … I think that’s what you call a webinar. Our first episode is on March 30th, where we will ask a lot of leading questions and hopefully get folks thinking about how to approach AI as they talk about it within their businesses. We’ll follow that with several other episodes, the second in which we’ll do a little deeper dive into the technologies that make up AI, but more importantly, the use cases, the business problems that those address, and try to bring some coherence, provide value through knowledge for our audience.

Julia:

For easy access, I will for sure link the registration form in the transcript below.