Hey, this is Jennifer Tribe and you’re listening to Frankly IT, episode 81.

Artificial intelligence and machine learning aren’t simply IT buzzwords anymore.

If you look at the Gartner Hype Cycle for 2020, you’ll see that machine learning is well past what they call the Peak of Inflated Expectations and is moving down into the Trough of Disillusionment on its way to the Plateau of Productivity.

The Gartner Hype Cycle, despite all its funny names, is a great way to understand the lifecycle of a technology and where you might see innovation and early adopters and where something is expected to become more mainstream. The Plateau of Productivity is really that fully mainstream status.

And for machine learning, Gartner is calling for it to hit the plateau in 2 to 5 years—that’s starting as early as next year, 2022.

AI is a little further back in the cycle but it too is past the Peak of Inflated Expectations and Gartner is also calling for it to hit the Plateau of Productivity in just 2 to 5 years.

So the time is now to be looking at how to incorporate AI and ML into your tech strategy. And that’s exactly what my guess, Sean Muller, says as well. In fact, he says if AI and ML aren’t part of your strategy today, you might already be behind.

Sean Muller is an enterprise architect and innovation technology consultant working with businesses of every size, from the very small to the very big. Sean is from the US but for the last 7 years has been based in New Zealand, where is also the host of the Kiwi Innovators Podcast. And he’s super passionate in particular about artificial intelligence and machine learning, and how they can help deliver on business goals.

Let’s hear what he has to say.

Listen here


Interview With Sean Muller, Innovation Technology Consultant

Sean Muller, innovation technology consultant
Sean Muller

Jennifer: Now, artificial intelligence and machine learning, AI and ML, have been buzzwords on everyone’s lips for a number of years. But like all emerging technologies, it was for a long time really only the big enterprise companies that were implementing or investing in it in any meaningful way. Would you agree with that statement?

Sean: Yes, absolutely.

Jennifer: OK, but now we’re getting to the point, or maybe I should say we are at the point where AI and ML technologies are within reach of businesses of any size. And you actually have a great story about a convenience store owner that illustrates that point so nicely. So why don’t you tell us that story?

Sean: Yeah, absolutely. And I want to make it even stronger than“it’s within reach.” My belief is that even in the small and medium business area, if they’re not looking at or using AI / ML, their competitors are and they will get disrupted. For an example of this, I, at a whim, was in a convenience store in New Zealand. We call it a dairy. And I was having a discussion with the owner-operator of that dairy. I mean, he owns it for his family and he runs it himself. And he told me he was using a machine learning algorithm to assess his stock, the stuff he had on the shelves, to predict what was going to sell and not sell over the coming months. Because he had all of the data. He’d been saving it for years and years and years. He knew exactly what he had ordered, what had been sold month on month.

Sean: And he had had a conversation with a consultant. And the consultant said, hey, you might want to take a look at that data and see if there’s some way that you can predict what you need to order. And he was able to cut his back room stock essentially by 20%. And for a small convenience store owner, the profit margin is very low. And so maintaining stock in the back room of the convenience store is a very expensive overhead. And if he doesn’t sell it in a timely manner, it becomes a problem for him. So he had cut his restock costs by 20 percent just by running this machine learning algorithm. And he said that it was getting better and better. He’d been running it for about six to eight months and he was actually predicting farther and farther in the future.

Sean: And so he could predict more stable stock on his shelves. And he had actually been able to restructure his shelves so he could optimize it for the stuff that was going to sell very quickly while still having stock that would sell occasionally on it. He had seen a great improvement within his business around this. And he was it. He was the whole company.

Jennifer: That’s amazing. What inspires literally a one-man shop to think, hey, what’s this machine learning thing and how can I use it?

Sean: The inspiration, the driving force sometimes is reading about some larger company that does something, seeing a television show or a Netflix show that talks about it. My advice would be if you are not thinking about it—and it might just be thinking about it—but if you’re not thinking about it, having a conversation with people about it, then you need to. Even small and medium businesses need to have roadmaps of where they’re going and where they’re going to be investing in.

Sean: But if we look back about five, 10 years. Those small and medium businesses were running everything local. I remember in the late ‘90s talking with a convenience store owner who had one of those Micro AS400s underneath his cabinet in his convenience store, where all of his stock and all of his financial processing, everything went through the little AS400. About five to 10 years ago, almost every small and medium business, they were the first ones to adopt cloud because their business outcomes and the technology they were using to deliver those business outcomes was small enough that it was very easy to pick up and use cloud.

Sean: That’s why they should start thinking about AI / ML now, because adopting AI / ML within the cloud, within the already deployed space that they are in cloud, whether it’s Azure or GCP or AWS, it doesn’t matter. There are machine learning capabilities within those cloud they can just subscribe to, test, see if it works and if they get a value out of it, they can continue using it, they can turn it on and off when they want to use it. And then if they don’t see a value from it, they can turn it completely off. And if the business owners are not thinking about this, let me assure you, their competitors are. And some of their competitors I mean, Wal-Mart is thinking about this. And so if you’re a small or medium retail company, you’re competing against a large enterprise already in Wal-Mart and they’re implementing machine learning that is going to make them even more efficient and capable to put you out of business.

Jennifer: It’s interesting that you mention the cloud, because this discussion about, this is where companies need to be. The big players are going there. You’re going to suffer if you don’t have it. It reminds me of previous conversations, first it was if you don’t have a website or if your business is not online, you’re going to fall behind. And then it was the cloud, as you mentioned, and now it’s AI and machine learning. So if you’re a company and you’re like, OK, I really need to look at this in 2021. Now, I know the shop owner, he was reading Forbes magazine and read this article, right. And so he decided to go out and find a consultant who could help him. Is that what companies need to do to find somebody who specializes in this, or is there a way that they could evaluate on their own where AI and ML might fit into their business strategy?

Sean: Oh, absolutely. Depending on what their business strategy is, how complex it is, what their roadmap looks like really determines the level they need to go to. And with the gig economy today, it doesn’t take very much. And the business owner doesn’t need to understand the coding of AI / ML, although if you’re going to start spending a lot of money on it, you really need to have a discussion with somebody. And a lot of money might be $10,000. You need to have a discussion with somebody that does know a little bit about it so that like anything else… I remember in the early days of cloud where we had outsourcing companies claiming to be clouds and selling basically outsourced hardware to small and medium businesses for exorbitant amounts of money and giving them nowhere near the availability numbers that true cloud could give them or the future proof capability of true cloud.

Sean: So, again, it depends on how much you want to spend. If nothing else, there are consulting firms that can come in and have discussions with them about AI and ML and a lot of them will do it… pro bono is not the right word, but they’ll come in and have the discussion. A good consulting company—and I’m not talking about the Big 4— a small consulting company can come in and can give an assessment or evaluation whether or not it makes sense to spend a little bit of money and see if it’s valuable. And by a little bit, I’m talking $2,000 to $5,000. Now, if the business is running on a profit margin that doesn’t allow a $2,000 to $5,000 investment in potential new technology that could fundamentally change their business model, then don’t do it. But if you’re in a small to medium business and you’re looking at your competitors and they’re going down this path, then it’s something I think you need to invest in.

Jennifer: Are we talking about custom engagements and custom scripting or could companies implement this through off the shelf purchases? I’m thinking back to an interview that I had on the show recently about service desk software and how machine learning is becoming table stakes to be incorporated into that software. So could part of your strategy being we’re going to invest in technologies that use this?

Sean: Yes, absolutely. This is actually a passionate area for me. Currently today in the large enterprise environments, only about 20% of machine learning or AI capability ever makes it into production. And it’s not because the machine learning or the AI isn’t successful. In fact, most of these machine learning models are very successful. They just don’t know how to get them into production because they work completely differently than any other software code that has ever been released in most of these organizations.

Sean: Where we see the disparity of that is in off the shelf products. Off the shelf customer relationship management tools, service desk tools, security software tools, operations tools. Those tools are now starting to come with AI / ML baked in. And the question becomes, how much money do you want to spend and what’s the business outcome you’re looking for? If there is an off the shelf product, that makes a lot of sense.

Sean: If you’re looking to spend under $5,000, there are very few off the shelf products that you can buy, but there are several off the shelf services that you could buy that might be able to give you the outcome you’re looking for. But again, standardized business outcomes. The age of cloud has done this to the point where businesses are now willing to accept service catalog items as opposed to custom engagements. And I think that’s the right way to go. I think an assessment of the business outcome and seeing if an off the shelf product or service gives them their outcome is a good step and it’s worth spending a little bit of time to research it.

Jennifer: We talked about how businesses need to be doing this because their competitors are and they’re going to fall behind. So is this a question of if you can implement it maybe this year you might have a competitive advantage over others in your industry for a short time? Or is it just at this point in 2021 you need to be there because you’re already behind?

Sean: Yeah, Jennifer, I’d hate to say it’s the second one. The reality is, is that in the United States and Europe and Australasia, if you haven’t already been thinking about this at this point, you’re actually a little bit behind. The larger enterprises that play in this space are buying up and putting out of business the ones that don’t. And the medium businesses that could see this coming… I’ve been talking to them across Australasia. I’ve been chatting with people out of the U.S. and Europe on a regular basis, the medium sized businesses—I’m talking sub $100,000 a year profit businesses—they see themselves being put out of business by—the pandemic set aside—by the enterprise now stepping into those spaces. Because that’s the thing with AI/ML is the enterprises can now do….they can do it your way, right? So it’s the Burger King mentality, right? We can do it your way.

Sean: The enterprise environments where previously before they were only aiming at the $10 million plus type clients can now aim at the little tiny competition and say, yeah, we could create a custom web page and we can put a custom AI chatbot and take over that one little niche that is five medium businesses across the United States.

Jennifer: Yeah, in the past, especially in the MSP space, there’s been talk about consolidation, but also talk about, Amazon, for example, they have a managed services division, but they will never take over the smaller companies who are dealing with people who like that one on one and very personalized interaction. But what I’m hearing is that with AI and ML, those big companies may be able to mimic those personalized connections and the smaller companies might actually be in trouble.

Sean: That’s absolutely right. The reality is, if you take a look at the Amazon grocery store in the northwest that was completely automated, was using AI / ML to basically you went in with your cart, you picked the stuff off the shelves and you walked out of the store and it just billed you. Amazon wasn’t planning to open grocery stores across the United States. They were developing a product that they could sell to grocery stores, that the grocery stores that buy it and again, this is an AI /ML off the shelf product, will be so wildly successful, it will put all the other grocery stores out of business. That’s the path we’re headed.

Sean: I’m based in New Zealand and I watch a lot of what’s been happening in the US during the pandemic. Here in New Zealand, we went to touchless pay for everything and stopped accepting cash anywhere during the lockdown and even the subsequent out of the lockdown. And the government is actually stepping in and making what we call payWave the mandatory paying purchase for credit cards. This is the new norm. If you can remove the friction for people buying, they’re going to buy more. And that’s even at the small and medium business level, that’s going to have a huge impact if somebody moves into the area that’s doing this low, low friction buying and they’re competing within comparable price to you, that’s going to be significant.

Jennifer: I’m imagining some people listening to this and being driven by the fear of being left behind or the fear of being put out of business. Are there some bright spots? Are there some reasons to get excited about AI and ML in a in a positive, proactive way?

Sean: You’re absolutely right. I was probably being a little doom and gloom. Anybody that’s been through sales training, oh, you’ve got to sell to the pain points. But the reality is, is that AI / ML functions can give real, real value. So let’s talk the context of the services that you were talking about.

Sean: Currently today, and I did some research on this for policing services around the world, most policing non-emergency services numbers are turning away 50 to 60% of the people calling in because they don’t have enough operators. And the ones that are getting through, about 60 to 75% are asking frequently asked questions that, yes, the person calling in could get the answer if they just want on the website and looked it up. But they didn’t. They picked up a phone and called. For anybody that’s ever called in and gotten an interactive voice response system, they’re terrible. Most people just start hitting zero when they get that because they want to talk to a person.

Sean: If instead the system said, what can I do for you? And when you said, what are your store hours? And the system came back and told you what the store hours is, that’s a very easy machine learning AI function. It’s a natural language processing. It’s speech to text, and text to speech. And it’s a known knowledge base that’s easy to have an answer for. If that takes 10 to 20% of your calls and handles them, that opens up your phone lines, your customer service lines to pick up 10 to 20% of the calls that people are hanging up and to give them better outcomes. And when they get better outcomes, they’re going to want to buy from me more.

Sean: From a customer service perspective, that’s a huge, huge upsell. These are base level functions. We’re not even talking, we haven’t mentioned deep learning. These are base machine learning / AI functions. This is not, we need to spend $100,000 and we need to build a system to be able to get these outcomes. This is in some cases off the shelf and in some cases it’s a small investment to be able to put something in place to be able to get these outcomes.

Jennifer: Any parting thoughts or, you know, main point that you’d want to leave people with?

Sean: I do consulting. So, you know, I’m telling you not to talk to a consultant all the time, but I do consulting and I do it across small and medium and large enterprises. One of the conversations I have with large enterprises is, is AI / ML part of their long-term strategy. And the interesting thing is, is that sometimes I get responses back of, “Well, technology is not a part of our strategy at all.” Technology needs to be part of an ongoing strategy that a small medium business has and definitely at the enterprise level. And AI / ML needs to be part of that strategy. It might just be this year I’m going to get on to cloud and I’m going to see how it’s going to work for me. And then next year I’m going to take a look at AI services, but it needs to be part of your strategy.

Jennifer: Excellent. Thank you so much for joining us today.

Sean: Jennifer, I really appreciate it. It’s been absolutely wonderful.

Links from this episode

Listen here

Like what you hear? Listen and subscribe.