Podcast

Full Episode Transcript:

Jon: Craig Le Clair, vice president and principal analyst at Forrester Research, and Stephen Siciliano, the partner PM director at Microsoft for Power Automate, join us today on the hello, Human podcast where we discuss the latest topics in artificial intelligence, and how it’s being applied in the real world. 

I’m Jon Knisley, the host of hello, Human, and a long time technologist working with companies to leverage first in class IT applications to win in the market. A big thanks to FortressIQ for sponsoring the program, and be sure to hit that subscribe button wherever you listen to podcasts.

In this episode, we’re going to explore how companies today are under near-constant pressure to accelerate growth, efficiency, and ROI. Effectively implementing intelligent automation at scale is crucial for success in the enterprise, but this remains a challenge with large organizations because of the lack of resources to streamline business process discovery, gather accurate data, and then easily kick off automated workflows. With these two true industry leaders joining us today, it promises to be an interesting and lively discussion.

Craig wrote a great book last year on the impact of AI and automation on the workforce. It’s called the Invisible Robots in the Quiet of the Night. I would encourage you to track it down. I’ll also put a link to it in the show notes. Stephen’s work deserves some congratulations as well, as the Power Automate platform has recently received significant recognition from the analyst community, which is just incredible given it just went into GA this spring. 

For today’s episode, Craig is going to start us off and give us a briefing on some of his new, not yet published research around the autonomous enterprise, which is fascinating. Then Stephen and I will join the conversation to discuss with Craig how technology in the post-pandemic environment is going to shape the digital workforce of the future. 

Welcome to the hello, Human podcast, Craig. I’m excited to hear the analyst viewpoint today since you have the benefit of talking to a diverse group of technology leaders and strategists and then get to work your magic and meld all that thinking together into coherent trends and outcomes. 

Craig: Yeah, thank you everybody for joining and also for that great introduction. Understanding today, Automate for tomorrow—that’s what we’re going to talk about. Thanks for the plug for the book. It’s doing quite well so I’m really thrilled about that. But let me get into the content here. I’ve been with Forrester for about 15 years. I’ve used that collective knowledge to try to bring all this automation together and what it means for our society and for the future of work.

Now, what we’ve seen in the last six months—we’ve all lived through it so I don’t need to talk about the effect, but I will in one small context—and that is this rate of acceleration in digital transformation. As analysts, we’ve been talking about transforming digitally since the mobile explosion in, say 2010, 2011. The iPhone had come out, and all of a sudden everyone needed a mobile app. The consumer had power in their pocket that the companies couldn’t deal with because they had gaps in their digital capabilities. And of course, we had disruptive companies with the sharing economy, mostly coming in and disrupting traditional businesses whether it’s the taxi cab business, publishing business, and so forth. 

Digital transformation was a topic and a strong initiative of the company, but very slow progress until we hit the pandemic, and now we’ve all developed these digital muscles. We’ve learned how to shop online. Obviously, we’re all working remotely, those of us fortunate enough to be able to. There’s been this surge in transformation and a search in intelligent automation, which I’ll describe briefly what that is. But it’s really been unprecedented, the amount of energy and investment that’s going into what I call practical automation. This would be task automation RPA, but also the more pragmatic AI building blocks. That combination is what we refer to as intelligent automation. Companies are basically redoing their road map for projects along the lines of this two-by-two.

Like all two-by-twos, it’s the upper right that is the more prized real estate. What you’re seeing there, what we call the acceleration zone, is automation that’s being applied in these areas, doing remote work. Our forecast,  as I’ll show you in a minute, is tens of millions that will remain working from home post-pandemic whenever that golden period is. Automation is needed to support them in different ways. A part of that is text analytics, which is being applied to paper processes, to documents, and emails, and forms. You can’t go to the office and shuffle papers anymore.

Remote business is another area, the number of banks, for example, that don’t have mobile deposits is pretty high like 30%, 40%. There is still a significant amount of business that depends on the face-to-face type of interaction, which is very difficult under the current conditions, so the answer is automation. The projects that companies had which might have been balanced between more tactical automation and more transformative AI-based business transformation, that’s the lower left, those projects are losing momentum. That emerging technology team that might have been working with open source tools, and trying to figure out how to instantiate the next-generation business algorithmically is now being de-emphasized for this more practical automation that has to occur, backed by the recession, as well as by the pandemic.

This is a rejiggered road map, if you will, that the buyers that we talked to, this is how they’re thinking about the world going forward. This is a bold attempt to try to define what the workforce looks like post-pandemic and where it’s been. Now, we call it the golden age of work. But for many years, we had relatively stable employment for a large percentage of the population. Automation expanded jobs rather than disrupted them more. You had an automatic nail gun, but the carpenter was still deciding where to put the nails. It just made that job easier. 

You had stronger labor unions, you had stronger and more stable middle-class benefits and income. But what you’ve moved to is a world where you’re going to see, with the recession and with automation, automation deficits, or job losses start to pile up, traditional safety nets are declining. You’re seeing social unrest. A lot of what we’re seeing on the streets today is this combination of factors. A part of it is this loss of hope for middle-class lives to be expanded the way they were in the golden age of work. There’s a lot of frustration underlying a lot of society right now and a lot of it has to do with work, a lot of it has to do with the pandemic, a lot of it has to do with anxiety-related automation.

The widening and exposing skill-gap that workers see when they go into the workplace and all of a sudden, the technology they’re dealing with is harder to understand, it’s maybe evolving at a faster rate than they are. These are all aspects of the new workplace. Tactically, digital elites spread far and wide. We’re going to see a significant growth in the talent economy that can be a human cloud if you will. We’re going to see a lot of workers are made at home, and so forth. Organizations will flatten and education is going to be rethought—the value of traditional education—with the pace of automation, and this development is going to be exposed as not being the most adequate way to do it. We’re going to need the private sector to step up and push education at work and lifetime learning.

A bit of a view of where we’re at this nexus is really an important point. I won’t go through this table in detail, but there is a lot of discussion of where we were pre-pandemic in terms of work from home and that’s that second column from the left. What we’ve done here is predicted based on the type of skills that you have in your work. On the left, you see these personas. Personas are essentially generic work categories where the occupations, of which there are 980 that the US Bureau of Labor Statistics tracks, we’ve mapped them into these smaller sets of generic categories based on these skills, thinking that automation will affect these skills in such a way that you can think of your workforce in these more simple and more generic terms.

If you’re a human touch worker, such as a massage therapist, you’re going to have this movement from 3% pre-pandemic to 8% where you’ll be working from home. You go down these areas and you see the most advanced knowledge worker that cross-domain. We think 35% of that population will remain working at home post-pandemic. All of this means it’s a sea change in how we’re working, where we’re working, that’s going to affect our lives, and basically, how we get work done. That’s a brief view of that.

Now, when you look at automation, this is a model that’s coming out and this is actually not published yet, but we have divided work in terms of trying to look at it in terms of how humans and machines interact. This is the basis of the pyramid. At pattern one, the software is completely controlling the process. We’ve been doing that for many decades. This workflow, digital process automation, deterministic, human-machine interaction that is defined and fully articulated by a task map, may be built with a process orchestration engine. 

The employee-driven is the interesting one for Fortress and for the intelligent automation market because here you see humans controlling the automation, firing off an automation. Basically giving instructions to a bot to go off and do something. This is an area that is going to accelerate in the next few years. It’s been accelerating right now due to the pandemic and I’ll blow that out a little bit for you. 

Semi-autonomous, you have humans and machines as peers. Your machine makes a decision or makes a set of recommendations, and provides them to the human. The human will make a decision based on that. You start to move at levels three and four patterns into this non-deterministic world based on AI, which basically is using machine learning and refreshed data. Some of that data can reformat the algorithms within the model. You get into the issues around explainability, you get into issues of predictability, you get into issues of emasculation of the human worker.

If you look at these four patterns, of course, four, completely autonomous is where the machine does everything. The human is not really in the loop in that sense, but the key is that to understand where automation is going in the future and how it can be best deployed in your organization, you need process technology to understand how the humans interact with the machine—what is the specific interaction, what are the implications of that interaction. You need to think about the pattern level that you’re at. This is not a progression. You can have a level one pattern that’s interacting with a level four pattern. When you recognize the characteristics of the pattern, you need to understand the right level of process understanding and process analytics that you need to drive the best result.

Now, as you move up those levels, you need to focus more on some of these aspects if you look at these seven aspects. These all get to be more important as you move from pattern one to pattern two, pattern three, and pattern four. Obviously, there is no real algorithm in level one. Level one is just a task-to-task type of deterministic workflow orchestration. But as you start to move to that level four that’s a closed-loop pattern, the algorithm has to be perfect. Otherwise, you’re going to have a bad result and this is where you need those kinds of checkpoints in the development of the algorithm, and testing it, and looking for bias, and so forth.

As you move to higher patterns, you need specific steps for bot mastery—for how humans will control the bot, how they will train the bot, how the bot will inform the human about ways that the human can act better. There is an interaction with humans from a learning standpoint with the machine. You need design thinking and digital worker analytics, and this is what we talked about in the last slide. The more you start looking at level two patterns, in particular, level three patterns, level four patterns, you have to have a way to study how humans are interacting with the machine.

You need code development because a lot of the more sophisticated algorithms have to stem from knowledge of the business and intimacy with the business. You need business, but also because you’re dealing with more advanced technology, you’re dealing with machine learning, you’re dealing with text analytics, you’re dealing with conversational intelligence, you need strong and deep professional development and data science capabilities. You need this converged development environment to collaborative low-code and deep development teams working together. That’s very important at higher-level patterns. And of course, it’s all about data, higher quality data is needed. The influence of data in an ongoing refresh, the basis of making decisions, that are predictive is used for decision management in the upper-level patterns. You don’t have that at the lower level patterns and of course, the governance for the reasons I cited. 

Employee anxiety as we talked about in the future work model, is very important. We’re going to see that we need to start designing these systems at levels three and four, where we understand what the effect on the human is of different patterns within the machine work. We’re just seeing that understanding the anxiety due to skill gaps, understanding that being able to turn to monitor on and off might be critical in this pattern, but not in this pattern. That’s going to be an area of greater discussion going forward in an area of greater research.

All of these patterns are all about the process, understanding the process. The better you understand the process, the stronger your human-machine interaction is going to be. And to do that, we have to solve this process-gap issue. You see, digital worker analytics, think of this as emanating from understanding the human inputs and outputs relative to a machine that might be in recording that occurs in looking at what a human is doing on a machine that might be brought into an analytic environment where you’re coalescing and summarizing human patterns across 50 or 100 different workstations and looking for anomalies, looking for areas that can be automated. That’s an important piece of this.

You then have traditional process mining, which in Fortress has a different way of doing this, but traditional process mining, you’re going in to log data—you’re analyzing that to get a view of that end-to-end process, looking at it from empirical data coming from the running process, building a set of technology.

Of course you have customer journey analytics. This is looking at what the customer is doing as they move on their journey from the more persuasive-type of interactions at the front end of the journey to the more service-type of interaction once they’re a customer. Understanding what they’re trying to do, their context is another whole area of study and area of gathering channel information to understand that. You have, on the left, process analytics had come from a process orchestration that’s built. 

The issue in the industry and I think going forward is that we have these software capabilities. They’re platforms that tend to focus on one or maybe two of these four areas, but they’re not bringing it all together. What we really need for the future of process is to get metadata from all these environments, bring it into a data environment that can then be analyzed and understood with machine learning and other techniques to be able to drive the running process. We’re not there yet, but we’ll get there. So with that, I’m going to turn it back over and we’ll have a short discussion. 

Jon: Thank you, Craig. Thanks for the update and thanks for sharing that autonomous enterprise model which is not yet published. That’s great and it’s always helpful to get the analyst’s opinion. Because you guys are exposed to many different companies and industries, you tend to see patterns and trends before anyone else does in a lot of cases. I certainly appreciate that comment on the digital muscle. Like a lot of people in the COVID era, I haven’t missed too many meals these days. I’m going to tell my doctor next week that I just have digital muscle. I got to get used to the new normal. Anyway, I’ll let you catch your breath for a second and get Stephen into the conversation here.

So, Stephen, we’ve heard from Microsoft that you guys saw two years’ worth of digital transformation in two months in those early stages of the pandemic. Just curious, at that pace continued, can you provide the sunny stories from the front lines? Who’s doing good? Maybe not doing so good? Feel free to just give us your color on the situation these days. 

Stephen: Yeah, absolutely. I would say the good thing is, things have slowed down a little bit. Suddenly, we’re seeing customers settle into this new mode of operation, and that’s particularly important because at the beginning it was this mentality that maybe the pandemic was a little bit like a sprint, when in reality it’s much more like a marathon. People are going to be operating in this mode for quite a long period of time and it’s causing permanent changes to happen across industries and across organizations. 

In the first two months, there was definitely a huge surge in getting systems that were previously hadn’t been touched in years or decades to a place where they could interact digitally with the other pieces of organizations. Luckily, things are settling into a more sustainable way of operating right now, and that is what we’ve seen. Now, that doesn’t mean, though, that there isn’t this huge backlog and demand for digital transformation that’s being accelerated, but I don’t think the acceleration is accelerating anymore. I don’t know what you call that in physics terms—the velocity cubed.

But to the point that you raised as well of, are there organizations that have done it well, and the organizations that haven’t done it well—we’ve seen even from both internal to Microsoft, and publicly, that there is a very wide spectrum of some companies and organizations. Even as Microsoft, in the first couple of months, we were working with educational institutions, we were working with public health institutions who had to be able to rapidly respond to things like COVID testing. 

In fact today with the power platform, we have many examples where COVID testing is still happening, it’s still out there, and it’s being built at running both from local all the way up to national levels on the power platform in an automated way. That was something that just wouldn’t be sustainable to build from scratch using the tools of the yesteryear where you had to build writing code, all of those things from day zero. That’s an example of where things have gone well because they had to.

On the flip side, there definitely are places where government agencies otherwise have not done quite as well. Generally, that’s happened where there have been these legacy systems in place, and there hasn’t been the appetite to put an automation layer between what people are facing, and the back ends. 

There is this story some people might be familiar with, where for certain unemployment systems, they actually put out an article saying please stop automating this unemployment system because we can’t handle it in the back end. That’s definitely an example where if you’re not building the right pieces in between what the users are interacting with and the back end, you’re going to run into problems. We’re seeing less and less of that now, that people are getting used to this “new normal,” but it’s definitely something to be cognizant of across all different types of organizations. 

Jon: I like that concept of transitioning from the sprint to a marathon, and I was actually thinking of the same example, it was with the paycheck protection processing. Essentially, the technology wasn’t too good and was creating too much of a backlog, and they told people they had to stop using RPA to process all those documents and contracts that were coming through. 

Craig gave us a good look into this potentially post-pandemic workforce over the next 10 years or so. What’s Microsoft’s view on changes to the workforce over the next decade, and the impact that technology will have? Earlier this month, Microsoft’s now not planning to reopen offices until July of next year, and 50% of work can be done remotely moving forward. What do you see coming for the workforce of the future, and especially in that 5-10 year range, beyond the initial crisis response, but not out in the science fiction land?

Stephen: It is interesting. For Microsoft ourselves, one thing to call out is we actually are seeing different velocities in different parts of the world. For the United States, for example, that was that July date, but there are other parts of the world that are doing much better. They’re already back in the office. What we’ll see are definitely different speeds that organizations are going back to working in offices, we’ll also see different proportions, and this is exactly what Craig was talking about. As you go down that list, there are some industries where you’re going to get to 8% remote work. But in other industries, you’re going to get 35%. 

What is particularly important is as people are working remotely, how do they have the tools to be successful in that environment. That’s something that will require in some cases up-leveling the digital literacy of the workforce to a point where they can be more familiar with things.

At Microsoft, we often talk about Microsoft Teams, but we all know that’s not the only video conferencing software out there—today, we’re using GoToMeeting. But up-leveling the digital literacy of the workforce is going to be really essential. If half the people are working from home, they have to be able to accomplish, not just get by but to really thrive in that digital environment, or else we’re going to lose a massive amount of productivity from the world economy. Finding ways that organizations can thrive digitally and that the workers can thrive digitally is going to be really essential. 

At Microsoft, one of the things we talk about is this idea of what we call the Remote Work Platform, which is built on top of Microsoft Teams and leverages the power platform to be able to give people ways to build out automation themselves, to be able to give people ways to use the chat interfaces inside of Teams to be more productive with their organization. The workforce overtime will transform from just being a consumer of platforms like Teams, of GoToMeeting, or whatever it may be, to also bring their knowledge of the processes that they have to transform them to be more efficient because you have to be more efficient with everybody working remotely.

Jon: That’s great, and obviously, this Work Remote opens up a whole nother issue around compliance, and how do you understand exactly what’s going on with your workforce that’s now spread out all over the world, but that’s probably a topic for another day as well.

Turning back to Craig a little bit, looking at that human-agent teaming, and the autonomous enterprise model, again, you walked us through those four patterns of interaction—the human in control, the machine in control, those at each end of the spectrum seem pretty straightforward. It’s that semi-autonomous layer where humans and machines share authority in that gray area which is a bit more curious, can you provide a bit more color, any examples of it in practice? My mind keeps going back to Matthew Broderick in WarGames, and the struggle between AI for human-AI versus human, and who’s going to be in control? That reference is probably too young for Stephen, I would guess, but some others in the audience might have picked it up.

Craig: Unfortunately not for me. You’ve picked on the one that is in some sense, the most interesting, and also maybe the most poorly defined at this point. But here’s an example, one of the areas that intelligent automation is jumping on is document-centric text analytics. We call it intelligent document extraction—basically going into forms, emails, and documents using natural language processing and maybe some surface automation to extract clean sets of data. 

The machine will do that, and maybe present confidence levels for fields that are extracted. If the confidence level, in terms of its accuracy, is below a certain threshold, it routes the situation to a human, and the human makes the final call on the accuracy of that extraction. That’s an example where it’s semi-autonomous, the machine is doing things that the human doesn’t understand how it’s being done, but the human is in the loop to share that authority.

The machines are doing a lot of decision management in that it’s saying, these 27 fields that I’ve extracted are perfect, and I’m not even going to show those to a human because I’m smarter than that human. But these three, I need to just send off to have another look at. That’s one where humans and machines are sharing authority, it’s not a closed loop. The machine is not taking that clean set of data and then determining the sentiment based on that coming from a customer, and then handing it off to somebody to deal with that sentiment, it’s bringing a human into sharing that authority. 

That’s an example that I think is good, but that pattern three is the one that has the most opportunity for humans and machines to do exciting things together. It’s your intelligence augmentation aspect there, a lot of the level two patterns that the RPA market’s so enthralled with are relatively simple task automation that you want a human to direct. But they’re not doing anything a whole lot different than what was being done before. They’re just listing out the human inefficiency. But level three and four, but for three gives you a really interesting way to think about the process in a very different way. 

Jon: Building on that a bit, Stephen, AI has been described as “infinite interns,” and having this immense but really non-expert capacity. When I first got involved in AI, 7-8 years ago, we used to say, if I can teach an intern to do it, I can teach the computer to do it, but if I require a PhD to do the work, I can’t teach the computer to do it. Given the advancements in the past 5, 6, 7 years, that’s starting to be a bit shortsighted. What’s your view on whether you can really combine the power of humans and machines through intelligent automation? 

Stephen: The example that Craig just laid out is actually perfect, and we can do that today with the technology that we have inside of Microsoft. One of the things that’s really important about enabling the infinite interns to do something is that you have an experience that can be used to train that AI, but that doesn’t require a whole lot of expertise in how the AI works in the back end, and understanding how machine learning works. Those types of things are very challenging concepts. If the person who’s trying to set up that process has to understand the intricacies of natural language processing, they’re never going to be successful in developing the models that are required to actually execute the process.

One of the things that we’ve been very passionate about is not only do we have the AI chops under the covers, but also surfacing AI through experiences that can be very easy to build, and get started with. Microsoft has talked about this idea of responsible AI principles. One of those key principles is transparency. That way you understand what the AI is doing, at least in terms of the output, even if you don’t understand the inner workings of every hyperparameter that goes into that particular AI calculation.

What it comes down to is how well can we build experiences to train the AI. That way, the humans who are actually part of that process can trust it because they understand what it’s outputting. As Craig mentioned, there’s some level of confidence that it’s always returning, so as long as we’re transparent with that, then you can have really any type of process that’s built out being used by people building out the process, leveraging that pool of “interns.”

Jon: That’s great. Obviously, that’s an issue of AI ethics and transparency is massive, and again probably a great topic for another session as well. How can we overcome this anxiety, given it will be difficult to re-skill workers in the short term. Stephen, you want to take that?

Stephen: Yeah. One of the things that we’ve been focused on is how do we lower the barrier to entry for automation. That way the people who are today doing those tasks can build out the automation and up-level their own skills themselves. This dovetails with the question that we were talking about a while ago is, how is the workforce going to change? One of the ways that the workforce is going to change is the people who are simply just executing that business process today, will be able to infuse automation as a part of it. As we think even about this pyramid that we have on the screen, it doesn’t remove humans from the loop entirely until you get all the way to the tippy top of that pyramid in the fourth layer of the pyramid. People are still a core part of the process.

From an anxiety point of view, as long as there’s still some need for that human intelligence, the decision-making capabilities that we have over machines—I don’t trust those infinite interns to make the core decisions of our business—you are going to be able to still have those combinations of humans and the automation that humans build working together. That will alleviate some of the anxiety that folks are facing.

Jon: Craig, you talked a bit about the process gap and desire of companies to address it. To me, this suggests this lack of detailed operational insights is a missed opportunity for companies, and if used effectively, could create a competitive advantage. What do you see as some of the top areas across the enterprise that could benefit?

Craig: I think the whole thing here that gives you optimism about getting a better handle on the operational processes is the fact that you can transition from the long consultant lead—and I love consultants—but the length of time and the length of the project when you have your traditional document thesis with human consultants doing interviews, and surveys, and so forth, and moving that to grabbing data transparently from these processes. If you see the digital worker analytics on the process binding, that’s what they’re doing. They’re actually gathering data and empirical data from the running process and applying advanced analytics to that.

That’s really the key here to get a more efficient insight, to get better insight, and get it more efficiently. You look at the ability to access data from 1000 workstation interactions across 100 humans and be able to synthesize that, and machine learning is brilliant at pattern, manipulation, and understanding, it’s perfect. You compare that to having to go out and try to understand what those people are doing. It’s just that technology allows you so much opportunity to get a better understanding of the operational process and ultimately to direct that process in real-time through orchestration and through automation. The ability to light up an algorithm that’s in the cloud to make a decision or light up a task automation digital worker to go do something based on what’s going on, that’s the real value of where I see this going.

A lot of companies are starting with the business shared service areas because there’s a lot of inefficiency in those areas. If you look at finance and accounting, if you look at HR, if you look at sourcing and supply chain and procurement, and these areas that every company has, those are the older systems in a company. They’re not the ones generating revenue, they’re not the ones that have been part of big transformations, there’s a lot of inefficiencies there. A lot of the acceleration of intelligent automation has been focused on clean extraction of low-value tasks and labor out of those areas. All of this lends itself towards the exploration of those processes. But that’s only the beginning because the line of business, the real processes that run companies ultimately can benefit tremendously from all this technology. I’ll stop there.

Jon: Cool and I’ll let Stephen wrap things up for us here. Once that process-gap has been addressed, it becomes then an issue of process orchestration, and managing, and executing those activities based on the data. What can we expect in this area in the next few years?

Stephen: You’re going to see a lot of organizations moving into this space. Even in just the past few months, there’s been a lot of focus on process automation from some of the bigger players in the market as well—Amazon has Honeycode, Google released their application platform, Celonis bought Integromat. There’s going to be a lot of options available to companies once they understand what the gap is, how they automate that, and how they build the processes on top.

Obviously, I’m a little biased because I work for Microsoft. We have a really great tool that folks can take advantage of, but you’re going to see a lot of different options and a lot of activity in the space. In the end, the competition would be great for all of us because it’s going to be able to expose all of the different ways that folks can be more productive, that they can up-level their skills and their capabilities in these very challenging times that all of us are facing. I’m excited about the upcoming capabilities, and all the things that will be built in the market in the next few years.

Jon: That’s great, and that’s a great point to end on.

To recap today’s episode of hello, Human with guest Craig Le Clair from Forrester Research and Stephen Siciliano from the Power Automate team at Microsoft. We were fortunate to get the latest thinking from two real industry leaders. With the ongoing return to business, effectively implementing intelligent automation at scale is crucial for success in the enterprise. However, this remains a challenge with large organizations because of a lack of resources to streamline business process discovery, gather accurate data, and then easily kick off automated workflows. By combining the strength of FortressIQ’s process intelligence and Microsoft’s Power Automate, organizations can quickly detail their current state operations, pinpoint the most optimal areas for automation, and execute RPA and automation initiatives on the fly with minimal coding required.

That’s a wrap on today’s show. Thank you, Craig and Stephen, for joining me, and for FortressIQ’s sponsorship. I’m Jon Knisley, and this has been hello, Human.