EPISODE #20: Applying Ancient Greek Wisdom to 21st Century Emerging Technologies

Guy Nadivi
17 min readApr 17, 2020

( Click here to listen to this podcast episode )

Over two millennia ago, the famous ancient Greek historian Thucydides wrote about how best to dispense with Spartans. He could not have known that 2,400 years later his writings would enter the pantheon of organizational thinking about innovation. Yet as unlikely as it sounds, that’s exactly what Mark Campbell believes has happened. As Chief Innovation Officer of Trace3, Mark uses the age-old reflections of Thucydides to help advise IT executives today. Navigating the labyrinth of emerging technologies is a Herculean task, and Mark has lots of sage advice on what innovations to take advantage of, as well as which ones to avoid.

In this episode, we chat with Mark about a number of emerging technologies from automation, AI, and machine learning to quantum computing. Along the way we’ll learn what questions a vendor should answer to ascertain if their product’s AI capabilities are based on engineering or marketing hype, the potential pitfalls awaiting any enterprise that decides to handle the “people problem” later, and the one biggest fear customers have about emerging technologies.

Guy Nadivi: Welcome, everyone. My name is Guy Nadivi, and I’m the host of Intelligent Automation Radio. Our guest on today’s episode is Mark Campbell, Chief Innovation Officer of Trace3, an emerging technology consulting firm based out of Irvine, California. As Mark’s LinkedIn profile states, he specializes in, “Squeezing the hype out of emerging tech.” Thanks to his work in the venture and startup ecosystems to identify emerging enterprise IT technologies and innovation trends, and introduce these to customers, partners and the industry. Mark and his team review over a thousand tech startups each year, and he’s also a frequent speaker and presenter on innovation, so we’ve invited him to come on our show to help us squeeze the hype out of such technologies as automation, artificial intelligence, and machine learning. Mark, welcome to Intelligent Automation Radio.

Mark Campbell: Well, thanks for having me, Guy. Greatly appreciate it.

Guy Nadivi: Mark, as a Chief Innovation Officer, you have an interesting definition of innovation, involving something called “positive deviance.” Can you elaborate on that a bit and how you incorporate it into your definition of innovation?

Mark Campbell: Sure. We get that question quite a bit from customers, especially customers starting up their own innovation group, or tackling an innovation project, or trying to inject emerging technology into an existing business model. The term positive deviance, I’d love to claim credit for it. It is so smart, but it was actually invented by a guy by the name of Dr. Jeff Degraff out of the University of Michigan, and then all of the very wordy and academic definitions that I’ve encountered over the years, I do think that Dr. Degraff has distilled this down into two words, the real essence of what it means to innovate — positive deviance, both sides of that. Kind of the idea there being that for every innovation, regardless of industry, technology, outcome, or pitfalls you’re trying to avoid, it means that you have to deviate from the status quo.

Mark Campbell: You have to kind of overcome your own internal business inertia, your own internal processes, your own internal way of doing business, your own internal technical skills. That deviation, of course, can take many different forms. Not all of them are beneficial. Sometimes deviation for the sake of deviation has gotten companies into trouble. Certainly, New Coke is a terrific example of that, but positive deviation where you’re kind of taking a look at this perfect future that you’re aiming for, and what sort of deviations are going to be the ones that give you the greatest probability of gain and the greatest probability of avoiding the pitfalls that comes from deviating from your current business models.

Mark Campbell: I think smart automation is certainly one of those examples that we are seeing in the market today, where automation is a terrific and wonderful thing that is helping us, if you will, streamline the status quo. When we talk about smart automation, we are really talking about deviating from that, and so I think that’s a very good example of how positive deviance, an example of it anyway, how we’re seeing it applied in our customer base.

Guy Nadivi: You’ve talked about there being three key drivers of innovation — fear, honor and interest. How should IT professionals factor those key drivers into their decision-making when considering moving forward with a potential enterprise innovation?

Mark Campbell: Well, I think before even jumping into an innovation project, or looking for innovation targets, or forming an innovation team, I do think the executive sponsor of the innovative initiative needs to take a look at these three core values, this fear, honor, and interest. If you prefer a little bit more modern terms, you can say fear, pride, and greed. This was a pattern discovered by a Greek researcher, oh, about 2,400 years ago, by the name of Thucydides. Now, at the time, of course, he was trying to figure out a more effective way of killing Spartans, but nonetheless, it actually applies very apropos today. When we take a look at this deviance, we’re going to deviate.

Mark Campbell: It does require us to evaluate the fear, the fear of staying where we are and having our competition beat us, versus the fear of changing, or potentially the pride and honor of damaging our brand with a failed innovation initiative, or increasing our leadership, our industry prowess by creating a new, innovative and disruptive product that changes our whole marketplace. I think like when we get looking at the interest or the greed, or what’s in it for us side, certainly there is a danger in releasing new innovation projects by displacing existing revenue streams, or existing products, or existing customer bases, and that has to be balanced against the potential upside of a new line of business, a new market, a new expansion, a new threat to bring against your competitors. I think innovation leaders, right at the very get-go, when they’re starting to contemplate using innovation as a weapon need to balance that, right? “What do we fear more? Where’s the greater honor, and what’s in our best interest? Doing this or not doing this?”

Guy Nadivi: Mark, let’s talk about AI and machine learning. Please tell our listeners where you would squeeze the hype out of these emerging technologies.

Mark Campbell: Well, certainly AI has had a long and roller coaster-ed history going back to the 40’s and 50’s, and in that period, we’ve seen kind of this ebb and flow of various techniques and technologies that are enabling AI to tackle harder and harder problems. However, when we take a look at a lot of products hitting the market, we trifurcate these into three buckets. We call them the simple, the savvy and the smart. Simple being just regular, procedural type algorithms, whether that’s a Excel spreadsheet or an autopilot on an airplane. Then, there are savvy products.

Mark Campbell: These are those that have embedded knowledge bases or access to expert systems. Of course, these were very popular in the late 80’s and 90’s, but really, they embody the knowledge of whoever created the product. Then, we get into the last of the three buckets, the smart bucket, and this is where solutions really learn. They take and ingest data, discover patterns, discover behaviors, maybe they discover baselines and report on anomalies from that baseline. Nonetheless, they are smart, they do learn and they do adapt.

Mark Campbell: When we look at this AI space, we are seeing hype being introduced by simple and savvy products out there, that are having their marketing department inject terms like deep learning, or machine learning, or AI, or convolutional networks, or reinforcement learning, or what have you, and to where their AI actually exists in their marketing department, not in their engineering department. That, that disconnect there can be very confusing for our customers who see a great story, they hear a great speech, they may even see a good canned demo, but looking under the hood a little bit, there really isn’t AI there. This is just AI-washing, a savvy or even worse, simple product. There are some techniques that we discuss with our customers on how to make that differentiation. One of the core truths about AI as we mentioned is that it learns.

Mark Campbell: It’s smart, and that learning is based upon a lot of data. A technique that we talk to our customers about is dig into that. When you’re evaluating a product and you want to really make sure that it is a smart product, not just the savvy product, talk about the learning. “How was this trained? How does it learn? Is it delivered in a pre-trained fashion, or does it continue to learn after I install it in my environment?”

Mark Campbell: “What data is being used for the learning? How much data is required? Is it canned data, is it publicly available data, or is it my proprietary data?” Digging into that layer of it when you’re confronting a potential smart solution. If the product out there does not truly incorporate any learning or any AI, you’re going to get very evasive answers, “Well, that’s a trade secret,” or, “Well, that’s a lot of smarts that our guys in the back room have injected into the product,” or, “Well, I’m not really able to go into that. I’d have to shoot you.”

Mark Campbell: If you keep pressing on that and get these evasive answers, you should kind of flag that as something really to be concerned about. On the flip side, if you talk to a true AI company, and you ask them, “Well, how does your product learn? What kind of data? Can I use my data? What happens after I install it? How do I re-baseline?”, you’re going to see them light up.

Mark Campbell: The analogy I use is like sitting next to a grandmother on the airplane, and you ask, “Do you have any grandkids?” If they don’t, they’re going to tell you, “Shut up and don’t be that guy,” and you’re going to get a silent plane trip for the rest of the way. If in fact they are a grandmother, you’re soon going to see Josh’s kindergarten play, you’re going to see a photo album, you’re going to see the birthday card that they got them last year, and you’re going to have a very, very conversation-filled journey. The very same is true with an AI product. If it truly has AI digging into that, it’s going to just open up a whole world, to the point that you almost don’t care what the answer is, but that enthusiasm and that passion that you see coming from the vendor, you can make a safe bet that you’re on the right path.

Guy Nadivi: What about automation? Where does the hype need to be squeezed out there?

Mark Campbell: Well, right now, we’re seeing a ton of automation products come to market. Certainly, what we talked about before about separating the savvy from the smart, equally true on automation products, especially those touting to be smart automation, so those all hold true. The other point of hype that we do see quite a bit in automation is the promise that this is going to be a single click and your problems are solved. Certainly, from a technology point of view, there are some great advancements out there. Certainly, there are a lot of techniques and products that truly will, in a smart way automate your business processes or your internal development life cycle or what have you.

Mark Campbell: The one thing however that is very often glossed over is the technology part’s the easy part. The cultural part is where things get a bit difficult, and these kind of happen on three levels. On a personal level, you do have people that maybe fear that automation is going to displace them, or at least displace some of the skills they’ve garnered over the years. At an organizational level, when you start talking about automating techniques within your group, there also is going to be a little bit of dissonance. Typically, organizations have well-worn processes.

Mark Campbell: They have rules of thumb, and certainly, if the automation is truly smart, it may suggest ways of doing things that are not part of the playbook so to speak, and that causes some organizational tension. The other thing that tends to happen at a corporate level is sometimes automation isn’t isolated into one particular team. Typically, when you’re automating, especially business processes, these start to leak over into other business units, other parts of the organization, and the cultural, political and personal ramifications of that, unless they’re addressed right upfront in a project. Even if the technology is perfect and flawless out of the box, these are some potential pitfalls that await to any enterprise that decides to handle the people problem later.

Guy Nadivi: What are you seeing as some of the most interesting innovations right now around automation, AI, and machine learning?

Mark Campbell: Well, we have a distinct advantage in that we’re partnered up with a few dozen of the world’s top-tier venture capital firms, so we do get an opportunity to see a lot of products when they are at the proverbial “two guys & a PowerPoint stage”, and watch them mature. Now, by the way, there’s a ton of infant mortality, and a lot of these companies don’t ever see the light of day.

Nonetheless, when we start watching these, as you mentioned earlier, we do have the opportunity to take a look at thousands of startups a year, you do start to see patterns forming, and certainly, when we look at areas that the venture community right now is spending a lot of attention on, certainly the AIOps, applying AI to IT operations. Smart SecOps, this is applying AI into security operations. Those two are huge right now.

Mark Campbell: There is such a large market out there, and there is such a dearth of products to satisfy that market that there are some very good products coming to market right now that solve a myriad of problems, but one other area is robotic process automation. We are seeing … As you’ve probably noticed, there are several products on the market that have IPO’d and their IPOs are enjoying a terrific ride right now, but that’s echoed in our customer base. When we go and talk about robotic process automation, whether it’s actually workflow automation, whether it’s screen automation, smart chatbots, call center interactions, across the board, we are seeing a big interest right now from our customers to bring those processes under automation and if you’re going to go through all of that smart automation. It’s not just about business process management or business process automation anymore, we are kind of seeing this, let’s say the maturation of the use cases that are being solved with AI, now allowing all three of those AIOps, smart SecOps, and robotic process automation to baseline and report on anomalies.

Mark Campbell: Sometimes this is called behavior analytics, to actually correlate, especially in the security space where you have thousands of alarms to correlate those down into clumps, and then for each clump, determine a root cause. We’re also seeing smart automation being used to not just react or control existing situations, but to actually make predictive alerting on things that could be going wrong, or bottlenecks that may be appearing, or issues that may manifest themselves further on down the line. At the very hairy edge in the automation space, and this is a little bit controversial right now, certainly in the security space, is automated remediation. If we are being attacked or we do have a storage array that goes offline, or we do have a workflow that all of a sudden halts, do we want automation to jump in and automatically remediate that? Of course, the answer is it depends.

Mark Campbell: Certainly, if it’s a low-level, we have someone from accounting that can’t get in because their password is jammed up, certainly stepping in automatically and remediating that, resetting their password, probably not that big of a deal, but taking an auto manufacturer’s assembly line offline, that’s a fairly financially onerous decision to make. I think over time, that’ll move, but that’s certainly the areas we’re seeing investment being made in today.

Guy Nadivi: I understand you’re doing a lot of exploratory research on quantum computing right now. How do you think quantum computing will disrupt automation, AI, machine learning for IT in the future?

Mark Campbell: Well, it’s still a little bit nascent, but we do have customers that are spending time and money evaluating quantum. Right now, the two hot areas are quantum computing, which includes quantum computing as a service, so instead of buying a quantum computer, just renting time on an existing one. That’s one big area. The other area is quantum encryption, and so that certainly leaks over into the security side of the house, but these are still in development. There are some great systems out there.

Mark Campbell: There are real products that you can buy today. There are open source projects that can be implemented today, and the main targets that these are approaching, one is optimization problems. These are your typical traveling salesman, flow dynamic, scheduling and network optimization. Not necessarily physical networks, but even human and social network optimization. Quantum is quite effective at solving optimization problems, even the primitive machines we have available to us today.

Mark Campbell: Certainly, when we take a look at automation and we’re talking about automating a workflow, today, what we’re doing is we’re actually automating heuristics. In a general sense on large scale processes and flows, it isn’t mathematically possible to come up on a digital computer with all of the combinations and select the best. However, with a quantum computer, that does appear to be a very solvable problem. As I mentioned, we do have small and primitive systems today, but even on medium-sized problems, that is becoming a little bit more of a reality today. This idea that quantum computers in the optimization space, at least, will be able to replace the heuristics being used in automation.

Mark Campbell: That definitely is a fairly likely outcome. The other area is AI training. You certainly can look at the training of an AI system as a non-deterministic and even probabilistic activity that once an AI system is trained, you’re not truly guaranteed that that was the optimal training. It just works with the training data that we’ve presented it with so far. There are…the term being bandied around right now is quantum intelligence, to where you can actually use a quantum system to take, again today relatively small AI networks and come up with the optimal training that is out there with a fairly high confidence. As these quantum computing systems mature and incorporate more and more cubits, the sample space of data’s going to increase. The amount of solution space that you’re able to address is also going to increase. I think that’s going to have a direct impact on smart automation both on the automation side and the smart side of them.

Guy Nadivi: Is there a single metric other than ROI perhaps that will cause you to recommend a particular innovation to your customers and partners over others?

Mark Campbell: Well, I think when we take a look at our customers, the one thing was, especially in the emerging technology space that’s a big fear is, “Is this going to be around tomorrow? If we implement this really cutting edge solution from a bunch of smart folks that have their own little startup, what’s the story going to be in six months? Are they able to keep up that trajectory? Are they still going to be around?” It really breaks down into this, “What is that product sale’s pipeline?”

Mark Campbell: Sometimes that’s a bit hard to measure, and, “How innovative is that solution? Is it the right type of innovation for the right time for the right problem, and how is the market responding to it?” Now, I know that I cheated a little bit and gave you three answers to that, but if you roll all of those up, it’s what we call momentum. When we take a look at a startup, certainly there’s a ton of other ancillary attributes that a startup has to have, like smart and experienced leadership, a great product suite, some good early results from their Alphas and Betas, but if you want to boil down one thing, it’s very easy to go look if a top-tier VC has already funded them. Now, if they’re onto their B or C round funding, that typically means that they’ve convinced at least two or three top VCs to do their funding, and one of the key attributes VCs look at before they write the big checks is exactly this momentum area.

Mark Campbell: If you don’t have access to funding data from VCs, there are a handful of emerging tech research companies out there that attempt at least to combine these. One example would be CB Insights. They put together something called the Mosaic Score, which is composed of market, the market strength that they’re targeting, the momentum they’re seeing in that, and how much money they’ve garnered, and how far have they burnt through it. There are metrics out there, but it all hinges around this momentum idea.

Guy Nadivi: Mark, what can CIOs, CTOs and other IT executives start doing right now to prepare for the innovations you think will be the biggest disruptors to IT in the next three to five years?

Mark Campbell: Well, I think that’s a very good question, and certainly one that we get brought in to deal with, and I think every customer realizes their market, their business, their culture, their skills, their budget are all very unique and shape that, but if I was to condense those down, I would actually put things into two buckets. The first bucket is what I would call defensive IT. This is using emerging technology to shore up your IT assets, set another way from a business point of view. This is to do cost reductions, efficiencies, to where the business isn’t worried necessarily about the money they’re pumping into their IT’s infrastructure, and the return that they’re getting from this. Typically, defensive IT helps buoy up those “ility’s”, availability, scalability, agility, portability, maintainability, a lot of those non-functional type requirements.

Mark Campbell: These are what we kind of call defensive IT. We’ve seen a ton of great innovations come on the defensive side, certainly things like containers, or cloud, AI, where it’s allowing us to do more with the budgets we have, or in some cases, even less budget. Being that defensive side, making sure that you’re doing the tried and true, as best and efficiently as possible, and increasing those ilities, I think that’s job one. However, if you’ve gotten to the point where your business has progressed by that, and I do mean the entire enterprise, has progressed from viewing IT as just a cost center, and therefore, the least cost, the better, now we start getting into offensive IT, and this is where IT starts getting a seat at the table for business decisions. We have an expression at Trace3 where we say, “All possibilities lay in technology,” and we truly believe that.

Mark Campbell: There are a ton of business problems out there, a ton of competition problems, ton of market problems, some regulatory problems, some skillsets problems, some budgetary problems that companies face, and we believe that there is a solution in technology for each one of those. The IT organization that consistently finds those fields, and brings in business, benefit from those, is going to be asked for a seat at the table. They’re going to be part of the thought leadership at the company, especially as it relates to lines of business, so not just the goodies that sit inside of the data center, but the actual lines of business and revenue streams and P&Ls of the company. That internal thought leadership, of course is much easier said than done, and there’s an awful lot of trust that has to be built. There’s a little bit of risk-taking that needs to be built, and just like we said before, a great evaluation of fear, honor and interest need to go into that.

Guy Nadivi: Offensive versus defensive IT, that’s a phrase I think is really going to resonate with the sports-minded CIOs out there.

Mark Campbell: Well, there’s a bunch of them. I totally agree.

Guy Nadivi: All right. Looks like that’s all the time we have for on this episode of Intelligent Automation Radio. Mark, it’s been great having you on the show to squeeze some hype out of the emerging technologies we hear so much about these days. Thanks for coming on.

Mark Campbell: Well, thanks for having me, Guy, and I’m certainly looking forward to your upcoming podcast topics. I think this is a terrific and fertile ground to plow.

Guy Nadivi: Mark Campbell, Chief Innovation Officer of Trace3, an emerging technology consulting firm based out of Irvine, California. Thank you for listening, everyone, and remember, don’t hesitate, automate.


Chief Innovation Officer of Trace3

Mark Campbell is the Chief Innovation Officer at Trace3 where he combines insights from leading venture firms and more than 25 years of real-world IT experience to help enterprises discover, vet, and adopt emerging technologies. His ‘from the trenches’ perspective gives Mark the material for his frequent articles and speaking engagements.

Mark Campbell can be found at:

Office: (303) 575–2144

Mobile: (719) 338–7772

LinkedIn: https://www.linkedin.com/in/mark-campbell-256748/

Twitter: https://twitter.com/HypeSnyper