Designing the Future of Work w/Donald Sweeney & Marshall Sied
Today’s guests — Donald Sweeney & Marshall Sied
Whatever it is, the way you tell your story online can make all the difference.
Listen to the full episode :
“We founded Ashling Partners to focus on moving people into more meaningful work by leveraging intelligent automation and digital transformation. Automation should be deployed to enhance the experience of both customers and employees, not just to reduce costs.”
In today's episode, I had the pleasure of speaking with Donald Sweeney and Marshall Sied, Co-Founder & Co-CEOs of Ashling Partners, the largest boutique hyper-automation professional services firm focused on driving business outcomes through automation, artificial intelligence, and process intelligence.
We talked about the startup and founding story of Ashling Partners and how Don & Marshall met each other. We discussed the impact of Process Mining on enterprise process improvement and how Process Mining now has a distinguished Magic Quadrant for itself, being a testament to the definition and separation of the field from the broader automation market. We went deeper into the promise of citizen development and stories of how it can be successful, the adoption and scale of generative AI technologies, and their impact on the enterprise. The importance of measuring outcomes and how to measure them through improved Employee (EX) and Customer Experiences (CX), along with getting executive sponsorship to increase the adoption of automation. We touched on how the future of work will revolve around creating more meaningful work for people as mundane tasks and processes get automated and digitized.
Donald Sweeney has spent his career working in process improvement and automation. Upon graduation from the University of Michigan, Sweeney joined Andersen Consulting (Accenture). In 1996 he joined the professional services group of the software company PeopleSoft, focusing on process automation and ERP projects. In 1998, at the age of 25, he started his own firm, Emerging Solutions, focusing on enterprise applications and process improvement. He sold that company in 2011 to Emtec and ran that group until his departure in 2017. He is the Co-founder & Co-CEO of Ashling Partners, which focuses on process automation and artificial intelligence to improve internal processes in organizations and impact the future of work.
Marshall Sied has 15 years of experience in helping organizations use technology to improve business performance and promote positive, innovative changes to their workforce. He is the Co-Founder and Co-CEO of Ashling Partners, a technology-enabled services firm that drives efficiency gains and business process improvement through Robotic Process Automation (RPA), Machine Learning, and targeted Artificial Intelligence.
Prior to Ashling Partners, Marshall spent time at Gartner working across strategy and advisory services to help Fortune 150 clients achieve digital ambitions through strategic planning, emerging technology and new operating models. Marshall has a BA in Business & Marketing from Michigan State University. He has completed a number of certifications in RPA, Big Data, Microsoft technology stack, and data storage. Marshall is a board member of the Illinois Technology Association's AI Council, and has written several documents on the impact of Digital Workforce on the Future of Work. He is co-author of the 'New API: Advanced Process Intelligence' framework and has spoken at several industry events focused on Robotic Process Automation, Intelligent Automation, and the Future of Work in the era of ML and AI.
“It’s important to align automation programs with the company’s overarching corporate objectives to avoid siloed efforts.”
“Process mining is really that discovery phase, right? You want to understand your process, where it’s working, where it’s not, and where there’s friction.”
Some questions we discussed:
Founding Story of Ashling Partners and defining a manifesto that cultivates talent: Don & Marshall share their unique and inspiring founders’ story that led them to the path of finding Ashling Partners. Even before starting the company, they wrote what is known as the “Ashling Manifesto”, which highlighted the key principles that govern the people at Ashling to continuously improve and become a better version of themselves.
Process Mining, creating a frictionless enterprise: Proces Mining by revealing process bottlenecks and efficiencies and allowing to create of an end-to-end process experience is the key component to drive process improvement, optimization, and automation. By getting its own magic-quadrant, it shows a testament to the market that Process Mining is just at the beginning of its growth journey.
The impact of automation on businesses: Automation helps businesses increase efficiency, lower costs, and optimize processes. It allows employees to focus on higher-value tasks, leading to improved job satisfaction and productivity.
The importance of aligning automation with corporate objectives: There is a strong need to ensure automation projects align with a company's broader goals. This approach helps maximize the value and benefits of automation rather than just focusing on short-term cost savings.
The right time to automate: Companies should consider automation when the potential benefits outweigh the costs and risks associated with implementing new technologies. This includes considering the impact on employee engagement and job satisfaction, as well as the overall customer experience.
Balancing automation and human touch: We highlighted the importance of preserving human empathy and interaction in customer experiences while implementing automation. Automation should enhance the customer experience rather than solely focusing on cutting costs or reducing human involvement.
The role of executive leadership: The responsibility for ensuring a successful automation program falls on executive leadership. Leaders need to have the right perspective and focus on the correct KPIs and metrics, prioritizing long-term value creation and customer satisfaction.
The role and promise of Citizen Development Programs: Citizen Development programs require an executive buy-in from the C-Suite to be effective, not only to have them build task automation but allow them to create robust solutions that alleviate business problems with tangible benefits and ROI.
Enhancing employee and customer experience: The ultimate goal of automation should be to improve employee and customer experiences. This means not only saving time but also delivering greater value and fostering positive interactions that lead to increased customer loyalty and satisfaction.
Transcript
Alp Uguray, Host: Welcome to the Masters of Automation podcast series. In today's episode, we have Donald Sweeney and Marshall Sied, the co-founders and co-CEOs of Ashling Partners, a technology firm focused on bringing hyperautomation to enterprises to help them achieve meaningful outcomes and enhance customer and employee experiences. So welcome, this episode is something that I long waited to have. So I'm very excited to have you guys join.
Donald Sweeney: It's been our life goal to be on your podcast, Alp, so thanks for having us.
Alp Uguray, Host: Yeah. Thank you, I'm honored that you guys are here. So to kick things off, Ashling Partners is obviously the largest hyperautomation consultancy firm in North America. And you both established a company that has a very great culture and a great way to impact the enterprises all around the world by adopting hyperautomation technology stack. But I think everyone wonders how you met, what you were up to before, what led you to start Ashling and define the values, and especially the Ashling manifesto as well, I think that influenced a lot of the people to feel inspired to join the company. So to kick things off, let's start with Don. What were you up to before, and then what led you to kickstart Ashling? And how did you meet with Marshall?
Donald Sweeney: Yeah, great question. Marshall responded to an ad in the newspaper. No, so Marshall and I used to work together, actually. So both of our backgrounds are actually pretty similar. We've both spent the majority of our careers in process improvement, kind of precursors to process automation, really focusing on business processes, and business process improvement. All of my career was really initially in that ERP space and then kind of expanded into an enterprise application consulting broader footprint. But Marshall and I worked together basically, the short version of a longer story there at our last company. And what we really found was implementing accounting systems implementing CRM systems, implementing budgeting and planning systems, all although great, we're very challenged actually to have that tangible, measurable ROI. And the other struggle that we had was if you implement an accounting system, the person who really doesn't benefit in many ways from that new accounting system is the AP clerk. Like they're actually pretty against the newest system. They had their way, they kind of got it down and they didn't really want something new. So as we were talking, we started talking about this future of work and kind of moving away from data entry, data manipulation, more towards more meaningful work. And at that point in time, we had sold our company to a larger organization. I had stuck around for some integration activity. Marshall had left and he was working at Gartner at the time. I think I'm allowed to say that, but so he was kind of on that, if you call it forefront of where the large companies were implementing kind of the newer technologies. And we would still get together and talk about what we would do next and all these kinds of things, usually over a beverage at a local bar or something like that. And he was saying, hey, this RPA thing, boy, a lot of companies are really leaning pretty heavy into this. This was the same time that we were talking about trying to find something slightly more meaningful. So literally on the back of a cocktail napkin, we started Ashling Partners. Ashling is the Gaelic word for vision or dream. And our theory there is that that vision that clients have had and companies have had for literally decades, trying to truly have efficient end-to-end process automation or just efficient processes can now come together, not just with that new ERP system or new, whatever system you're talking about, but with this kind of consolidation of optical character recognition, machine learning models, your BPM now called IDP, you're basically you're long running workflows, your RPA, like all of these are now getting stronger, better, the cost of technology has come down significantly, the speed and improvement of technology has improved significantly. So that vision of finally having that fully end-to-end efficient process, we felt was finally coming to fruition.
Alp Uguray, Host: And Marshall, I think Don covered pretty much your guys' story. But is there anything that you can add there?
Marshall Sied: I can comment, Alp, that this will probably be your longest episode yet. So whether you're planning on that or not, it's gonna be a very long episode. I love it. No, I wouldn't add much to that. You know, the only other thing I would say since you asked about, just the founding of Ashling, is you basically start, we started Ashling with what we wanted the company to be, and also, just as importantly, what we didn't want it to be. And so, you mentioned the manifesto. We crafted that before we even incorporated the company. And we found that a lot of the value system, the principles we put into that manifesto, tended to be turned externally with our clients. So, our business has changed. We need people that are versatilist and willing to be, the Darwinism approach, willing to adapt quickly, because, change is happening with faster increments nowadays, with emerging tech. I mean, everybody's, under the next wave of exhaustive speed-to-outcome technology with OpenAI and ChatGPT and large language models, right? And so, we kind of embraced that, and we wanted folks in a nucleus and a culture that embraced that. What we have found is that clients also want that. So, when you connect that to, more meaningful work, the future of work, and some of the challenges from, the macro economy, from scarcity of talent and labor in the market, it all kind of aligned and made a lot of sense. You know, I wouldn't encourage everybody to do it that way. It just, it worked for us to think about what we wanted from an internal perspective, even before we looked external, which, we knew we wanted to be in this intelligent automation space at that point. But, that's the only thing I would add.
Alp Uguray, Host: Yeah, that's very interesting. And Don mentioned that the accounting software that is built for AP clerks are the ones who were the most resistant to adopting the new software. And I think that's one of the highlights of the future of work as well. Like the adoption part of it and defining a culture, both internally and externally, where people can develop the digital skill sets and adopt the software built for them is key. So based on that, and starting the Ashling, you guys wrote the manifesto, you had a very good understanding of process improvement, ERP, and the RPA, as well as the emergence of RPA, because it was also the early stages of it. How did you see the best ways to structure this customer adoption? And what are some of the things that influence the enterprises that are sold on stay innovative? And Marshall, do you want to go ahead first?
Marshall Sied: Yeah, I was going to hop right in on that one. So I think it's, some of the some of what I'm probably going to talk about here today is, buzzworthy. We hear it a lot in the marketplace, but I actually mean it. Joking, of course, but, I think being outcome obsessed really helps. So when we started the company, and we're kind of looking at the right service offerings, knowing that, once again, we need to, we're going to need to change them because the technology is going to get adopted departmental first, then cross-departmental, end-to-end processes, then it's going to go enterprise-wide, which is certainly what has occurred thus far. But we really started the company, really focused on outcomes as the North Star, which means you need to have an advisory function. So we started the company to be more management consulting-esque on the future of work and how automation is going to play a leading role in changing the type of work our human workforce does while also helping companies that adopt early and more broadly be more competitive. And we kind of got pulled into more of that focus. You know, that full life cycle, plan, build, run, measure and improve. So, starting with that outcome approach and a little more in a consultative way is certainly benefited because education is the key right now, and we're already proliferating. We started with RPA and now, if you believe Gartner's hyper-automation definition, I mean, there's eight capabilities that make up automation today and it's going to grow. And you're kind of seeing people that start to work in the lanes and start to work in the lanes. You know, we've had a lot of early-stage and early-stage automation companies that really stayed neatly into their lanes, start to swerve into other people's lanes slowly. You had RPA vendors start to swerve into the traditional OCR and IDP lanes. You saw them start to swerve into the process mining lanes. People are just swerving all over the place right now. There are no lanes. So we need to evolve that pretty quickly and make sure that we're even more so acting as a great aggregator and building those multi-year roadmaps across multiple technologies while consolidating from a total cost of ownership perspective where we can. I think combining those capabilities in an enterprise really benefits each other. Someone who actually leverages process mining ends up being able to build a pipeline for automation and process improvement opportunities.
Alp Uguray, Host: Don, you had expertise in process improvement as well, like in your previous companies. So like tying that to what Marshall said on building the advisory function and then the capabilities on top of that to allow enterprises to educate themselves better and innovate themselves. What were some of the themes that you thought companies are really not doing well and they could think differently or actually doing really well and should do more?
Donald Sweeney: Yeah, I think there are a couple of ways of doing that. I mean, one of the things that we try to do is leverage third-party benchmark data. So it's not us subjectively saying, hey, you're doing well, or you're not doing well. It's coming in and having them compare themselves to their peers, whether it's in the industry or the size of the organization. And so really walking in and trying to identify continuous process improvement metrics. Now, these conversations really work out well when a client has a continuous improvement mindset or a lean six sigma kind of mindset, then this is great. I mean, they don't have to have that, but if they do, then they totally get it. Because you're walking in really, first and foremost, trying to have business outcomes, as Marshall was talking about. And then secondarily having a tool conversation or a technology conversation. But the technology is really just to be the vehicle to get from point A to point B. You really want the conversations with the client to be focused on where are we trying to go? We're at point A, and we want to get to point B. What is point B? And to that point, I think starting from the outcome and then the solution and then the technology approach really helps them to frame how to solve their pain problems.
Alp Uguray, Host: I think some customers also get pulled into the hype a little bit, like, for example, there's the AI hype, a generative AI hype, and the blockchain hype at the time. And before that, it was RPA. And I think there are a lot of industry reports coming out on top of a lot of noise coming up early on. But now, RPA became intelligent automation, became hyper-automation. So to Marshall's point, that tech stack kept growing for an enterprise. So what are some of the things that you guys think will be like, maybe not five years from now, but three years from now, so that they can anticipate better?
Marshall Sied: So, number one, yeah, five years is probably too long of a crystal ball prediction now. That's number one. Three years is probably as far as anybody can go at this point, which brings me back to Darwinism. Everybody's just gotta be willing to adapt and kind of swim with the current. When you look at everything coming into the market today from a large language model perspective, I don't think we've had a client conversation in the last month where it didn't come up at some point. Until risk and governance are figured out in the enterprise, now it's great for creative skillsets, it's great for individuals, but at the enterprise level, I think it's going to take a little time from a governance perspective to understand the credible sources where the models are pulling from, to be willing to allow your people to put potentially confidential and compliant data back into a large language model like ChatGPT or BARD or whatever. And so I think you're going to start to see a lot of adaptive governance models, what we would call adaptive governance models, kind of start to take root in the next three years in order to actually take advantage of this powerful technology that is absolutely something that people should be investigating and trying to conCitizen Developer Programer how to leverage it better. If you look at how we think it'll actually proliferate in the enterprise, you're seeing all these enterprise application providers start to provide an integration service to OpenAI to whatever your large language model of preference is. So I think it will be the more embedded type of conversation as opposed to a standalone and pure replacement. Everybody always jumps to the peak of inflated expectations at the very onset of new technology, whether AI or automation. I remember when it was mobile, and then before mobile, it was packaged applications like ERP and CRM. But it tends just kind of to be more accretive as opposed to displacement. I mean, it happened in ERP where we had an ERP where you started with an accounting package, general ledger, AR, AP, and then you had a fixed asset module, then you added a supply chain management, or if you're a healthcare organization, materials management module, and then you added the whole HR Citizen Developer Programe of the house. And it just started to proliferate based on a technology stack, a kind of containerized packaged application with a UI, a good database layer, and the ability to customize from a coding and studio perspective. I mean, I don't see it being much different. I just think the change will happen quicker with what's happening right now in the space.
Alp Uguray, Host: And Don, what do you think about it?
Donald Sweeney: Yeah, I think we're very early in the ML AI journey, although Microsoft, with how much they've released in the last two weeks, makes us feel like we're pretty far along that journey. There's still a lot more to go. I mean, there's still so much data entry and data manipulation that people do. I think the question is, where do we see this in three years? I think it's that end-to-end automation, ML AI driving a lot of components of that being, in essence, the traffic cop of data and kind of moving the data through the process. But there are still people today who receive an invoice via email, an AP invoice via email, open up the email, open up the attachment, read the attachment, hit print, walk over to the printer, grab the invoice, walk back to their desk, log into the accounting system and enter the exact same information that's on that sheet of paper that was on the email in the first place, right? So connecting many of those things and having AI and ML start to improve maybe or streamline some of that process and then improving the customer experience, improving the employee experience of that process. There's still a long way to go in a lot of the day-to-day activities. So I think a lot of the plumbing is going to be improved over time. And what is very interesting is also the vendors setting the expectations right. So like in the, I think if they sell, and this applies to RPA process mining and any vendor out there that works with tech, sell that, oh, this is magic. If you buy this, you adopt it quickly, and it solves all their problems, and they all go away. And then half the journey, figure it out, oh, just cannot do that, or I cannot do this. And I think similar to maybe LLMs, there's that aspect as well, where there's so much unknown and so much hype. It's hard to do the reality check of what it can accomplish.
Alp Uguray, Host: So in your perspective, I think LLMs, standalone and RPA as well, standalone and based on what you've seen over the past five to six years and how people understand and set those expectations to change over time, what do you guys think about it? Marshall, do you want to go first with this one? I see you were nodding.
Marshall Sied: Yeah, regardless of the tech, whether it's RPA, LLM, or IDP process mining, the number one truth is businesses want outcomes. And once again, I warned you; I was gonna be a little cliche in this conversation. And I think what you see a lot of times is that some of the software providers, and I find them to be very, very instinctive and to listen to their customers well. So they develop products on their product roadmap and release features quicker than ever. They make acquisitions where it makes sense to fill some of the gaps in a voice they have along the automation journey if you will. But I think many software vendors end up stopping a little short of the value realized. So they end at the value theorized but don't get to the actual value realized. And it makes sense, right? They have to put software on the streets in the hands of customers in order to have the ability to impact value. But I think you're going to see a lot of organizations really focused on kind of the rollout strategy and trying to figure out a way to take something departmental or regional to global using reusable assets that you might've developed in North America and rolling it out to APAC and using that as a vehicle of change for the process as well, by the way. Hey, you have to do it this way if you want to leverage this type of automation. And here's the business impact it will have, by the way, so it's worth doing. So change your process, right? You kind of saw this a little bit with cloud applications. I remember when we started migrating Oracle PeopleSoft on-premise ERP clients to Oracle Cloud ERP, and you told them, hey, you can't customize in people code anymore. You have actually to adopt the cloud template that Oracle gives you, or SAP S4 HANA gives you. And it was an emotional conversation culturally for a lot of large organizations. They're like, well, everybody likes the idea of change until it happens to them. And so I think the fact that you kind of force people to think about change and to re-engineer their business a little bit, I think that's where people are right now with automation, where I think that the technology's there, the software vendors are there, the consolidation will continue to happen. Probably in that space, but we can't stop short of value. And I think you're going actually to see a lot of uplift in big value realized case studies, where up to this point, you've kind of seen big use cases, maybe not huge program value success stories, but you have a lot of programs that are kind of in that value theorized category. They're not to value realized quite yet. So maybe I'm a little biased because I'm a service provider in that space. But I see that to be a challenge for many clients. It's like, hey, we have this huge opportunity. How do we actually get from point A to point B so I can actually claim that value? Because that's the beauty of this technology. You can actually touch the tangibility of business value. You know, that's one of the values where we get really excited about things like process mining. I know we don't want to get too into the weeds here on different tools and technologies and stuff, but you think about process mining, it gives you an X-ray, if you will, of how a transaction flows through your system. And it does it for every single transaction. So, you run a hundred thousand transactions through the system, and maybe 2000 or 20,000 are running in one way. And then another 20,000 are running in a slightly different way. Instead of A to B to C to D, it goes A to X to Y to C to D. Or something like that. So, you really can start to quantifiably measure, hey, if this was standardized and automated, here are all the steps we would have removed from this process. And you can cost out those steps, or you can figure out speed to value, auditability cost reduction, fees, or whatever the case may be. There's usually a cost to be paid. And then from B to C, there's usually a quantifiable, measurable value, thereby tracking both kinds of the before and after effects of these activities.
Alp Uguray, Host: And I think that's very valuable, especially like I've seen recently in process mining have their own magic quadrant right now. They rank the vendors, which is a big improvement in the industry to call that out, to tie itself as a separate process improvement, diagnostics to define where the problems are at and then solve them. I mean, we could nerd out a little bit more about process mining, I think, right now. Well, what do you guys think about that? Them having their magic quadrant separately, and what does that... What does that really mean?
Donald Sweeney: Well, I think it shows you, if you follow the customer journey, in which everybody should have a customer centricity and kind of work backward from there, it shows you that one of the major challenges to overall business operations improvement, of which automation is but a component, a capability for business operation improvement. That process visibility is still a critical bottleneck. Whether it's through acquisitions or different localized process requirements from a regulatory or cultural perspective, it shows you that businesses don't fully understand their processes. And at the same point, they're not connecting those processes to true opportunities to improve through automation. It's still very siloed, but I think it just shows you that the customers are still very, you can't automate if you don't understand, right? That's actually a very dangerous proposition to start automating processes you don't understand for many reasons. Upstream impact, downstream impact, just bad customer experience, employee experience, supplier experience, depending on who your stakeholder for that process is. So to me, that's the big takeaway is that you're seeing a lot of people gravitate towards something like process mining because they don't understand their processes. And they don't even know where they should focus their automation engineering teams or process improvement Lean Six Sigma teams. That being said, I think there's a lot of upCitizen Developer Programe potential for process mining because what we hear from organizations is, hey, this kind of feels like data and analytics and process analytics and data mark for my order to cash process or procure to pay process. And they're not wrong by that. But they're also not thinking about it big enough where you have to take action orientation with outputs from process mining. Whether it's standardization, optimization, automation, or all three, frankly, if you're reimagining and doing some process clean sheeting, you can claim a lot of value that way, but you have to act. Right. And I think we're at a point where people understand the value of the potential of something like process mining or broader process intelligence, but it's not connected yet to act. The RPA, IDP, and low code, but we're getting there pretty quickly. And I think that's what really excites us about what we've tried to build at Ashling is that end-to-end experience. What do you think, though? You think about process mining, sometimes called process intelligence or execution management. You know, it's a lot of different terms, but. Going back to when Marshall and I worked together previously, you would go in and interview people for a couple of weeks to implement an ERP system. You'd sit down with, like, the accounts payable person just to keep with the same examples before. And you'd ask them, OK, how do you do accounts payable? And you would sit there, and you'd map that out. Well, lo and behold, you just spent a couple of weeks mapping all these things out and then start building it. And when you build it, it's not exactly the way that they told you it was. And now you're going to give them a change request because they didn't tell you about these extra three steps that happened only 20 percent of the time. They told you the process as it occurs 80 percent of the time. And that's not a bad job on their part. It's just kind of human nature to focus on the parts that they're aware of and maybe not give you everything. But that's, you know, that's a bad customer experience from a consulting services standpoint. And that's a risk in implementing something. Now, the beauty of process mining, or whatever we want to call it, is it gives you every transaction going through the system. So you can see the one percent outliers. You can see the 20 percent outliers. You can truly run this on an ongoing basis. I think where process mining is really in its infancy. Is people don't see it as the ongoing execution management again, that more lean six sigma approach of continuous improvement. They look at it as, OK, I built this, and now I see how my activities are, and now I'll automate those bad activities. You know, I look at process mining as something that you want to look at month over month, quarter over quarter, and make sure that you're continually improving and continually moving the needle, so to speak. On standardization within your group, cost improvement, cost efficiency, depending on what the group is, speed to value, and better customer experience. Again, whatever the process is, you're defining that success and then measuring it through your actual actions, the actual transactions flowing through the system. So if your activity is not getting better, well, then how do you think you're truly getting better on how you treat your customers just because you say so? Right. So I think this is going to be a really big growth area. I kind of see it as the brain of automation or the hub, central hub of automation and standardization and continuous improvement. And then all things kind of resonate out of this. So very, very bullish on the future of the area.
Alp Uguray, Host: And I think the magic quadrant really a testament to that belief and showcases the process mining will evolve itself as its own plane. I really like the aspect that currently, it's a little bit viewed as a diagnostics, but it helps you to view how the process functions, where the bottlenecks are, and where the pain problems are at. But I think once they get to live transaction processing in those systems, you get an alert on an outlier immediately, and maybe that triggers the automation or automatically changes or ping someone in the company to take action. At that stage, I think that end-to-end process innovation is a way. So which is really fascinating. I can talk about this a lot too. It's a really, really important aspect of it. So I like to take the discussion to something that's been evolving over time I think AI and automation are coming into people's job descriptions and then changing their roles. I think the skill set also changes over time. So like, for example, now people are expected to write great prompts to interact with an LLM. I think similar to being able to write great prompts is knowing how to ask a question to Google, and it gives us a search query. And I think from the other Citizen Developer Programe aspect of it is the automation, where someone interacts with the bot themselves to automate their tasks with humans in the loop. So as this evolves over time, now I'm coming to the question part. What do you guys think about more of the ethics of AI discussions, and what does that mean for an enterprise? I think there's a meaning for overall to be careful in general not to be judgmental in the way AI predicts, but from an enterprise perspective, what do you guys think about it? Marshall, do you want to go first with this one?
Marshall Sied: You give me the big philosophical question. You know, look, I think this is an evolving area to AI ethics, which is certainly something that I think will be a growing field. And when we're talking about fields that will actually grow and expand based on some of this emerging technology stuff that's coming into the landscape, I think ethics is a major one. You know, AI trust is another kind of topic you're hearing a lot about nowadays. I think people will start by understanding what data sets we will be feeding to an algorithm. And if it's financial and transactional data sets, I don't know how biased that can be. Frankly, at a certain point, it's not going to be that biased. It's when you bring it into hiring practices based on historical data sets. Hey, maybe we've always had biased hiring practices, for an example. And we're making our future AI algorithm even more biased because all we're doing is feeding it our own historical data set. So, I think there's going to have to be scrutiny on the data sets, which is why I think it's going to take a little time for something like a ChatGPT to really be used at a very macro enterprise-wide level, because you just can't predict the type of bias that has been fed to that large language model yet. I think it will actually kind of condense before it expands again. But I do think it will expand eventually. Right. So I think it's a question that nobody has an answer to. And if they say they do, they are probably not humble enough regarding truly think about all the tentacles of the implication that are out there. But we do need to have that as basically a stage game. I mean, you should think about that in testing. What is the data set if you're about to deploy a model? What is the business problem I'm trying to solve? Does that business problem have a lot of qualitative, subjective data on feeding it based on history? Do I think my history was biased? I mean, there's just it's based. There's not going to be an answer, but there does need to be an AI decision framework for you to deploy anything in the enterprise.
Alp Uguray, Host: So that's a very good point. What do you think, Don?
Donald Sweeney: Yeah, I think the AI decision framework is the key. Maybe my first thought is you could probably do a whole hour-long topic just on AI ethics. So there's another one for you. But. You know, we struggle with and people in general struggle with, OK, if the ML model is going to use the data that you gave it, OK, well, what's the bias in the data that you gave it? And so, by nature, people evolve and change over time. You know how I thought and acted two years ago is different than now, let alone 20 years ago. And if you are creating a model by nature, the model is static on its data. And it's not expecting somebody to change. It's expecting that data to be repeated. And so if X, Y, Z occurs, X, Y, Z is likely to occur again. Well, maybe, maybe not. You know, there's just so much loaded into AI ethics and bias and all these things. Language usage when you're looking at resumes, they talk about how people are selected in the AI models in viewing resumes based on the language used in the resume versus the school you went to or the grade point or the work experience or these other things. So there's just so much loaded in that question. But I think that's a huge area for growth and development. And I think we already know people coming out of college with that as a degree, AI ethics. And I think you'll see more and more people coming out. I think that is a growing field.
Marshall Sied: Yeah, the only other thing I'd add, just listening to Don out, is that there's another term AI explainability. You're starting to see that embedded in some of the data science and ML building platforms. And if you can explain why the conclusion was reached by an algorithm or why we have a certain percentage of confidence in our kind of output from an ML model, then you start to be able to think about putting more of this in production with confidence and with non-biased in mind. So AI explainability is a theme, a keyword on top of AI ethics, that I think you'll continue to hear more and more about over the next couple of years.
Donald Sweeney: Yeah, thanks, Marshall. I think I went off on a diatribe and didn't really get to my original point. That's the point you need to have transparency and trust in your AI models or it's not going anywhere. So knowing why you got the response you did and trusting the response because of that input is how AI will become pervasive across all of these areas.
Alp Uguray, Host: Yeah, that's very interesting, especially like a lot of the like. Even those who built these algorithms don't know why the output is that way. And that leads to the data that leads to, I think, your point about AI explainability of the concept. I've seen, I think, today or it was yesterday, now people are collecting signatures to say maybe, let's pause this movement a little bit and then maybe figure out the regulations and then maybe the explainability aspect before it scales to the rest of the world. This is, again, one of those things I think that can have its own episodes.
Donald Sweeney: Well, and again, if you're looking for episodes, I mean, ChatGPT, everybody's talking about it. And it's kind of the cocktail party conversation du jour right now. But two weeks ago, it was the solution to like every problem. And now, already only two weeks later, you've got corporations saying we can't use that because we don't know the data that went into it. We don't know what we're going to get out of it. And frankly, we may not want to share our proprietary data into that model because then the public has that data. So, right or wrong, you're already starting to get many people moving away from that due to the lack of transparency. So it just goes back to further Marshall's point.
Marshall Sied: I absolutely agree with that.
Alp Uguray, Host: And I think one aspect of it is right now the integration piece. And I will tie this to a little bit more on the enterprise process improvement. I think from the integration standpoint, and it's easy to integrate. So people adopt very quickly and then try to find an application use case. I think similarly for RPA vendors; for example, with the citizen development, they have a user interface that's easier to adopt for people to be able to build automation. And obviously, it doesn't function as near to AI because it's much more cognitive there. But I like to take the discussion to a bit of the movement of citizen development and your perspectives around how companies can start adopting it. So it's not only the core RPA, COE technical team who knows how the RPA function, but maybe citizen developers to assist with large process automation or task automation. So I heard some people say it doesn't work. Some people say it works great. What do you guys think? And then why do you think some people say it doesn't work? Where are they seeing the wrong, and where do you guys see it being the most successful? Marshall, do you want to go first?
Marshall Sied: Yeah, sure. Somewhere in between is my answer regarding those who say it doesn't work and those who think it's the next coming. So I don't think you can have a CITDEV program without a centralized program. And I don't think you should have a centralized program without at least Citizen Developer Programering a Citizen Developer Program program. And it depends on your organization's maturity and your operating model's structure. There are a lot of inputs you got to think about. But I think one of the challenges you may see with Citizen Developer Program is that I go back to my outcome statement. It's hard to measure business value when it's Alp creating something just for him based on how he does his daily job. You know, when he grabs his cup of coffee, he runs attended automation really quickly. And it's not scalable to everybody on Alp's team, everybody in Alp's department, and across the globe of Alp's company. So I think the business value is something that still needs to be addressed with Citizen Developer Program programs because they tend to be more groundswell, bottom-up programs. We view it as a spoke to the hub, a more centralized program where the centralized program should govern it. But, there has to be a degree of flexibility from executive leadership that is funding a Citizen Developer Program program to say, hey, in the near term, we might not have the same value drivers, the same value metrics that we expect out of our hub, our professional team. But we want to; we want to incubate this a little further. Our experience is that nine out of 10, quote and quote, named citizen developers drop out of a program. And it's because they have day jobs. So until you, until you truly get folks to have the ability throughout their work week actually to dedicate to that, you're probably going to see mixed results. That said, you want to have proactive programs and training academies. We've set up a lot of these academies. Our take on it is that you're probably going to see more citizen consumers rather than citizen developers. And that's what no code is. You've got, you've got your professional developers using low code tools to get to outcomes quicker via automation. And, based on componentizing that automation, it could be a drag-and-drop marketplace potential for a Citizen Developer Program to consume, depending on their persona, their role, their departments, their need, the applications they have access to. There's a lot that goes into it, but certainly I still feel like we're very much at the beginning stages of seeing full-blown Citizen Developer Program programs, really show quantifiable business value at this point. Yeah, I certainly agree. I don't know if I really have much to add from that. I think there will always be a need for corporate governance, corporate basically building of the Lego pieces if you will. And then, the citizen developers will largely assemble Lego pieces that make sense for them to have a bespoke automation or bespoke assistant or whatever to do their portion of their role. But there's still going to be a need for a lot of building those blocks and building up the governance and everything else if it's truly going to work. It's got to be simple and it's got to be valuable, right? If it's simple and valuable, citizen development works. If it's not either simple or it's not that valuable, it doesn't work. And with standards and frameworks in mind too, right? Because you have to think about long-term technical debt. I mean, somebody's got to support that. If it breaks, do you think the citizen developers are going to be the one that re-engineers that all the time? They're going to be calling somebody, right? So I think that's the CIO's fear of CIDF programs. They inherit that technical debt potentially. And if it's a lot of technical debt, that's something that they might not be prepared to inherit quite yet.
Alp Uguray, Host: The measurability, the value part is quite interesting as well because they're typically task automation and like me picking up my coffee or like starting a report and measuring that and driving, tying that to the full business is really, really gets tough to do. And to your point, the technical debt is also an interesting one. Then if someone is a citizen developer, is their Citizen Developer Program is a side gig, or is that part of their job? I think that's where things break.
Donald Sweeney: At some point, this matures into not needing to be a specific, measurable business value, but I think we're quite a ways away from that. So everybody needs Microsoft Excel, but you can't quantify if you didn't have Microsoft Excel, how many hours you would have to spend doing other things. It's just it's a part of your job at this point, right? Or Word or whatever, email, whatever example you want to use. At some point, I think a digital assistant and some kind of citizen development is just part of everybody's job, right? You now are going to take these three components and connect them together, so you don't have to do that every Monday manually, but it's just part of your job to do that. So whether you call it a Citizen Developer Program side gig or whether you just call it part of your day job, I think we will eventually get there. We're certainly not there and not there in the near term.
Alp Uguray, Host: So, one thing I wanted to ask you guys is that we talked about end-to-end experiences and changing the process to drive the most value. And we talk a lot about employee experiences, customer experiences, and there's the payer experience, the patient experience, and it keeps going on. And obviously, this speaks to its own industry and how the definition changes based on the industry. But how would you define an employee and a customer experience for an enterprise? Don, do you want to take this one first?
Donald Sweeney: So maybe, sorry, repeat the question?
Alp Uguray, Host: Yeah, so how would you define employee and customer experience as a benefit-driven by automation?
Donald Sweeney: Yeah, so I mean, there are hours back to the business as a benefit. Separate to that, there's the customer experience benefit. So people sometimes use the word frictionless customer experience, right? A lot of technology companies use the word frictionless in their buyer journey. So you want as positive of a customer experience as possible. That means making it simple to order from you. That means the Amazon experience where you suggest other things based on your previous purchase. Maybe even suggest when it's time to purchase again based on reviewing your purchasing trends. That's all part of the customer experience. Same thing with the employee experience. Employee experience now is about making sure that they get answers to their questions when it's convenient to them, not just when HR is working from 8 to 5 or something like that. Or getting their questions answered on their benefits and whatever other examples you want to have. So these are all going to become table stakes eventually. An employee is going to work where that is provided versus somewhere else where it's not provided. So I think it's all a measurement of success. There are different types of measurements of success when you're talking about automation and what's important. And that's one of the first questions you ask going back to the very, I think, first question about Ashling was driving the business outcomes. So that's why we sit down with our clients and first ask what's important around this business outcome. What are we trying to achieve? Marshall, what do you think? Things are out there—my keywords, which is usually frictionless. You got to do it when it's convenient, regardless of who your stakeholder is, employee, supplier, or customer.
Marshall Sied: The only other thing I would add is a contact center to give you an anecdote, right? I don't view it as I don't view automation being just about reducing the average call time. I mean, we all work so hard to get new customers and clients. So you don't want to automate it so the experience is worse or even neutral. You want it to make it better for the customer. And if it means shortening the call time or giving them a mobile automated workflow experience that is more convenient, great. But you shouldn't automate it down to the point. Your value drivers change is the point. That is a great time to service your current existing clients, which we all know is the cost to service a lot lower to your existing clients than to any new clients. So you should take advantage of that and wow them with the experience. And same goes for your employees and your patients and your nurses, and whatever industry you're in. But that would be the one piece because I think your value drivers have a different index. I think it changes slightly from a waiting perspective. And let's not take it. Let's not lose sight of the fact that it's not a cost out when it comes to customer experience. We've got to improve the experience through automation, not try to remove any human empathy from that experience.
Donald Sweeney: That's a great answer, Marshall. And I love how that kind of almost brings the conversation full circle where we started with why we started the firm and about moving people into more meaningful work and the future of work and all these kinds of things. You know what Marshall was just talking about there: if you're in a customer service area, it's not about solely cutting costs and driving efficiency. It's about empowering that person to do what's truly meaningful. And that might be spending some time with the customer. Right. It might not be about just getting off the phone. It might be about building a relationship with the customer and then automating all the other tactical activities they spend their time on. So if you could remove 70 percent of the administration and then that remaining 30 percent becomes the 70 percent where you're actually engaging with your clients, you're probably going to have a better client base or customer base. And the responsibility does squarely fall on executive leadership, though, to make that happen. Right. Because some of the KPIs we see out there right now is about cost out, which can be an element. You can deliver more effectively and more efficiently. But, just calling out that executive leadership needs to have that perspective is why we're very big on ensuring you're aligned with corporate objectives. Or else you might think you're not going to fall out of the park with your automation program, but you're frankly focused on the wrong KPIs and metrics. Using two examples, I won't use names in the first one, but how many people enjoy calling their cable company or utility? Right. It's usually a pretty painful experience because they've tried to automate and cut costs out of that experience. You got to sit there and hit your five, then your nine, then your two, and then enter your phone number and do all these things before you can actually talk to somebody. They've cut so much cost out. It's a negative experience. Yet, if you call Apple customer experience, it's a very positive experience. You kind of want to hang out with that person like they're super smart, knowledgeable, and friendly. I don't know, maybe it's me, but I kind of like hanging out with the Apple person. So that's a really positive customer experience. And therefore, they're a premium brand, reinforced by that customer experience. Nobody sees their utility or their cable company as a premium brand.
Alp Uguray, Host: Yeah, I think that really brings it home. Especially understanding how to design the program automation and drive value, that outcome is meant to be employee and customer experience. And that doesn't mean only time savings, but all means bringing great value and a successful outcome to your end customer who likes, becomes a repeat customer, or loves to chill at Apple stores like Don or me. I enjoy that as well.
Donald Sweeney: So that's... You'll see me as the old guy just in there playing on the iPad.
Alp Uguray, Host: I want to thank both of you for joining the call and sharing your perspectives, experiences, and opinions. I'm really honored to have you both. So I really appreciate this time. Thank you very much for joining.
Marshall Sied: Pleasure is all ours, Alp. Thank you so much.
Donald Sweeney: Thank you.
