Nischay Mittal, Global Head - Automation/AI, Zinnov, converses with Max Cheprasov, Former Chief Automation Officer, Dentsu, to unlock the intricacies of everything Automation. How can enterprises scale their Automation journeys to achieve the desired ROI? What is the playbook for enterprises to truly scale Automation practices? Max, a true-blue digital native, answers these and more such burning questions and sheds light on multiple interesting aspects of leveraging Automation to holistically transform enterprises and generate significant value in the long run. Tune into this podcast and take away interesting and insightful food for thought.
Nischay: Hello everyone and welcome to a brand-new episode of the Zinnov podcast Hyper Intelligence Automation series. I'm Nishchay, Principal and Global Head for Automation at Zinnov and I'll be your host for today.
Over the past one year, the Automation space has been humming with activity from massive fundraisers to IPOs and big-ticket acquisitions to Automation companies exploring newer use cases. We've witnessed it all. However, despite the exponential growth and opportunity in the Automation space, enterprises are struggling to scale their Automation initiatives. How can enterprises scale their automation journeys to achieve the desired ROI? What is a playbook for enterprises to truly scale automation practices? To answer some of these burning questions, we have with us Max Cheprasov, the former Chief Automation Officer at Dentsu.
A Stanford alumni, Max has successfully integrated his dual passions, technology and business through his education. A true-blue digital native, Max has closely witnessed the intense focus on all things Automation by enterprises and will shed light on some of the interesting aspects mentioned above. Welcome, Max. It's great to have you here with us today.
Max: Yes, thank you for inviting me. It's an absolute pleasure to be a guest on your podcast.
Nischay: Perfect. So, let's dive right in, Max. So, Max you’ve had a very exciting and diverse professional career for over 25 years. You've been a part of several fast-growing firms over the years. You’ve also founded a couple of start-ups, with one, in fact, even having a successful exit. And more recently, you were with Dentsu for almost 10 years, serving as the Chief Automation Officer. In fact, one among the only two Chief Automation Officers globally. Talk to us about your journey so far and how you've navigated this ever-evolving and dynamic technology industry, including your tryst with the automation space.
Max: Yeah, looking back, I've always been interested in finding new and better ways to service clients and enable employees to do their best work and this of course requires that people, processes, and technology, work in sync and in harmony with each other at all times. There is this Japanese management philosophy, Kaizen, which is the best way to describe what I was trying to achieve in my various roles.
Essential to that Kaizen methodology is that culture of continuous improvement, elimination of waste, and the importance of providing a great workplace and employee experience. And think of the simple formula you've seen in business, ‘Happy employees equals happy clients.’ So, a few years ago, I found myself at a moment in time when my teams were reaping the full benefits of applying Kaizen at scale and just when I started to feel that we're about to exhaust all of the traditional ways of achieving operational excellence, we began experimenting with applied Artificial Intelligence and Automation technologies. And in those early experiments, we used Natural Language Processing and API-based Automation...For example, to assist with simple admin and back office activities like scheduling meetings, reviewing legal agreement, then we did some more complex front office-related operations, like creating insightful reports for clients, automating service delivery workflows, and optimizing resource allocations. And the unexpected boost in terms of effectiveness and efficiencies that we saw convinced me, and of course the executive team that we're entering a new era where businesses that thrive and grow are the ones that will be leveraging the power of Artificial Intelligence across the entire business at scale and not just in any one area of the business.
So, in 2017 when I was the Chief Automation Officer at Dentsu and for context, Dentsu is the fifth largest media and advertising agency group in the world with over 65,000 professionals in 145 countries. So it's a big business. In 2017, I introduced a multi-year plan with a mission to elevate human potential, and then inject AI and Automation into all of the areas of the business.
And fast forward to today, Dentsu has over 400 Automation experts, over 300 AI-powered digital robots in over 500,000 hours of productivity returned to the business on an annual basis. And that might sound like a lot. But I think that all of us are still at the early stages of the Automation adoption curve. And according to a recent update from Deloitte Insights, for example, the percentage of businesses that have more than 50 automations in place reached double digits for the first time this year at 13%, and they've been doing the survey for five years now. So, there is still a long, long way to go before we see the first, truly and fully automated enterprise where every employee has an Intelligent Virtual Assistant at their side.
Nischay: Wonderful, Max. It's truly illuminating to learn about your journey so far and also on your perspectives around the Automation space. I especially liked your point around how people, process, and technology need to work in sync at all times and your focus on the popular Kaizen philosophy. And of course, how this has gradually transformed into AI-powered Intelligent Automation.
So, diving deeper into Automation, Max, there's always been this raging debate on how Automation will lead to job losses and this fear that embracing Automation would mean enterprises would have to cut down on FTEs in order to really drive cost savings. So, what's your reaction to this battle between automation versus job losses? And more importantly, how did you manage this apprehension back in the days when you were getting started with the Automation initiatives at Dentsu?
Max: Well, I'm happy to say that the mission of the Automation COE at Denstu since day one has always been focused on elevating human potential, not on eliminating it. And I remember that back when we established the COE in 2017, Forrester predicted that by 2027 Automation will displace some 25 million jobs, but only to add 15. And that loss of 10 million jobs in the US over a 10 year period due to AI was a conservative estimate back then, as I saw some predicted a net loss of up to 17 million jobs.
But today, a more encouraging update, for example, from this year's World Economic Forum is that by 2025, there will actually be a positive net gain of 12 million jobs due to AI. I think that reality is going to keep changing over time and will largely depend on what the businesses want to achieve through Automation and how quickly AI will continue to evolve. We're still far away from the strong form of AI, but the current form of available technologies can certainly replace certain professions, especially those that are exclusively dealing with transactional activities like data entry, document analysis, and system updates, for example.
But it will be a lot more challenging to replace professions that depend on higher cognitive abilities and collective social intelligence, like creative problem solving, critical thinking, and strategic planning, for example. And reflecting on the past three years at Dentsu, the conversation among the automation project sponsors has certainly shifted from cost cutting and FTE reduction to improving employee experience and well-being.
I think COVID-19 certainly contributed to that mind shift, but most importantly the Automation experts were finally able to provide real evidence that there are greater benefits to investing in people, teaching them new skills, and upgrading their work experience, essentially allowing them to be superhumans with the power of AI in their hands. And this human and AI collaboration has been found to be 85% more productive than either humans or robots working alone. Now, that's Kaizen on the AI storage to me.
Nischay: Great, and so the fact that Automation will, in fact, be a net job creator coupled with extensive change management is what you would recommend to alleviate this fear of job losses associated with Automation.
Max: A 100 percent.
Nischay: So Max, were there any other challenges you faced by setting up the Intelligent Automation COE at Dentsu? Or if I paraphrase this, looking back, what are the two things you would have done differently provided an option?
Max: Yeah, we faced many challenges when we set up the Automation COE at Dentsu. The greatest one of them all was the limited talent pool with the relevant skills and experience, and very little guidance and best practices or lessons that are in the main industry, in general. I mean, four or five years ago anyone embarking on the RPA and Intelligent Automation journey was certainly trailblazing their path forward, almost unguided. And in the marketing and advertising industry, we were the largest company attempting to use those technologies at scale.
So, one advice I always give when starting something new is to have a solid plan with a clear purpose in mind and our ultimate north star was to inject Artificial Intelligence into the operating DNA of our business. When we were getting started, there were limited examples of use cases, where and how AI can be effectively applied. So, we had to do a lot of experimentation and we had to spend a lot of time making very careful choices where we will invest our time and energy to ensure our initial projects provide the expected benefits and we don't disappoint anyone in the early days, especially if we take on too much.
But today, there are many well documented use cases and solutions out there they can immediately apply. Many system integrators and partners who can reliably help you get started quickly and at scale. But the other advice I always give, start small and be relevant. You need to find a small, but a meaningful improvement that you can make in the business using AI and demonstrate its value. Then quickly follow up with more use cases and large improvements. When we were highly selective about our initial projects, we surveyed our C-suite. What business challenges were top of mind for them and then we found a couple that had been lingering on everyone's housekeeping list for some time and we quickly got to work to resolve them, or as quickly as we could back then, but it took us several weeks to get the POC off the ground and today you can accomplish the same results in days, instead of weeks using low-code platforms. So, my advice, start small and start with Low Code/No Code tools to quickly show the value and benefits, then scale.
Nischay: Perfect. So, two great pieces of advice emerging for anyone who's starting new with their Automation initiatives. One, of course, have a solid plan with clear purpose. And the second is to start small and be relevant.
So digging slightly deeper into this whole COE aspect, Max, we have been doing a detailed analysis around the Fortune 250 enterprises and their Automation footprint, and an important and interesting facet has emerged. So as per our analysis, we found that almost 90% of the Fortune 250 have already started to invest meaningfully in Automation initiatives. However, at the same time, only around 40% of these enterprises have currently invested in an Automation-focused Center of excellence or COE. So given this fact, how critical is it for you to establish a COE in order to scale Automation deployments and what do you feel are the reasons the same?
Max: Yeah. Well, the good news is that I also saw another survey that almost 90% of those enterprises that do have a Center of Excellence in place are effective in delivering business value using those Automation technologies. And I always believe that in order for the business to get the most value and return from the AI and Automation, those technologies need to be integrated into every business function, horizontally and vertically across the entire business. And then every employee will need to have a virtual assistant helping them in everyday work. Essentially the business that wins in the future will be the one that will have AI in its operating DNA and if that's going to be the end goal, then the Automation COE is going to be critical for the successful adoption, evangelism, and democratization of AI and Automation across the business.
With that mission in mind, you will need to assemble a team, identify the right partners that will help you along the way and this may include technology platforms, implementation partners, and specialized recruitment forms. But most importantly, you will need to create a strategic roadmap and plan outlining everything from why you started the COE to how you plan to make progress, where to begin, where to continue, how much investments will be needed, how you will control and manage your bots, and how you will scale and report them the value this COE is delivering.
That's a bit more than three factors that you asked for, but I'm sure I missed a few things. The good news for anyone getting started today is that you now have access to lots of resources and lessons to learn from the early pioneers and you can accomplish the same results much faster today.
Nischay: Also Max, there are several operating models for a COE, including a centralized model, decentralized model, federated models, and so on. In your opinion, what is the best operating model that an enterprise should strive towards?
Max: Yeah, the quick answer, it depends. The decision to embrace either model...Yeah, it needs to be driven by the organization's unique circumstances and that strategic plan that you have and the purpose. I will personally always default to the federated model when it comes to recommending a model that works best when there is a large, globally distributed enterprise with highly decentralized decision making.
That worked really well for Dentsu in the very matrix global and complex environment with many semi-independent business units and P&L across 145 countries. The decentralized model would not scale in that kind of environment. It would've created a lot of duplication resulting in higher cost of Automation for us and the centralized model would have quickly turned into a bottleneck. So, the federated model allowed us to centrally define and enforce processes, establish policies and standards that everybody can follow, and have a more effective knowledge reuse and certainly made it a whole lot easier to achieve economies of scale.
Nischay: Perfect. So essentially the COE model decision depends on the organization and the vision. But in an ideal state, the federated model should work best even if you have a large globally distributed enterprise. So, one critical function of an Automation COE is around this whole process discovery. So how do you ensure a healthy pipeline of use cases from an automation perspective?
And again, there's been immense traction in the industry around a lot of these tools-based discovery methods, including Process Mining and Task Mining. However, as per one of our recently concluded surveys of around 257 global CXOs, we found that more than 90% of the CXOs continue to rely on crowdsourcing-based manual techniques for discovery. While at the same time, around 25% of these are also starting to experiment now and adopting Process Mining and Task Mining tools. So, what do you think is the reason for reliance on such manual techniques and what are the inherent advantages that some of these Process Mining and Task Mining tools provide?
Max: Well, I think you need to have a balanced approach, one that effectively leverages both, the Process Intelligence tools which are not cheap today, but you also need the crowdsourcing-based manual techniques, like the interactive workshops and surveys. And the manual techniques are cheaper and faster to deploy for initial discovery, but they mainly lead to a pipeline of high-level ideas and use cases that are based on mainly anecdotal evidence and incomplete information about the process. So, further analysis and deeper process discovery needs to be done with a subject matter expert anyways. Because without that information, how do you choose which use case to prioritize?
So, this additional deeper manual discovery process can take weeks, if not months before you have the complete information needed to guide the design of a solution and then it will take another few months to engineer and deploy a solution before you start to see any benefits. So, by then, you may have talked to a few dozen stakeholders, you got them all excited, you talk about Automation vision, you talked about their challenges and needs, but you only had enough resources to satisfy a couple of stakeholders. So, there's still a benefit to these manual techniques, like awareness building and generating internal support and buy-in, and creating a grassroots Automation movement across the business.
But I prefer to combine that with the AI-powered Process Mining techniques to automatically conduct a full time and motion study across all of the ideas, regardless of their potential size and the impact. And the Process Intelligence tools can capture all the inputs and outputs of the end to end process across all applications with all its task variations, and essentially capture the full truth down to the keystroke and mouse click. They can be expensive, but in my experience, if they're effectively deployed, they will pay for themselves 10 times over by saving you time in the cost of firing a small army of business analysts. So ultimately having access to a complete view of the processes end to end creates a huge opportunity to quickly make decisions on which use cases to prioritize without wasting your COE’s valuable time on manual process discovery or work. And this approach significantly reduces the time to market for each new Automation solution.
Nischay: Perfect. So, Max, there’s merit in adopting a somewhat balanced approach, as per you, that is almost like a hybrid approach of using crowdsourcing techniques coupled with AI-powered tools-based methodology, so that would help an enterprise to combine the best of both worlds and really reap the benefits of both these techniques to some extent.
Max: Yeah, that's exactly it.
Nischay: Perfect. Max, while process discovery is critical, this can at times generate 100s and 1000s of processes for Automation which is a good problem to have. But then the professional developers are not able to cater to all of these processes., especially the very low complexity ones. So, this is where the Low Code/No Code program becomes extremely strategic for enterprises. We know that you again had a very strong focus on Low Code/No Code back at Dentsu, so talk to our audience a bit about the Low Code/No Code phenomena and how this can be leveraged effectively. In fact, what, as per you, are the three best practices that an enterprise can essentially follow in order to have a successful Low Code/No Code strategy?
Max: Yeah. So last year we launched the Citizen Automation Program at Dentsu and upskilled over 100 employees in various business functions, including HR, Finance, Marketing, Sales, and IT, to use UiPath’s low-code to UX platform and that was an instant success. It allowed us to tackle our Automation pipeline from the bottom-up...Through those discovery conversations, and surveys, and workshops, and process discovery technologies, we discovered lots of opportunities. And so, this program allowed us to spread even more awareness of Automation as more people can actually get the tools in their hands and start automating. Just imagine, our two-day hackathon...a group of newly upskilled citizen developers would generate enough value to pay for the annual cost of licenses and training within days. So, it's a real thing and beyond improving productivity, the program creates other benefits for citizen developers and Automation consumers, like happier employees and higher customer satisfaction, for example. I think, it's really important to choose the right technology that will support your citizen development, goals and ambitions. The right technologies and best practices here are the first thing you need to prioritize when choosing how to start your program. Is it a truly Low Code/No Code technology? Does it meet the needs of your use cases, because you wouldn't want your citizen developers to encounter challenges with the platform when it can’t support their use cases and requires additional plugins, or you need to get additional help from a professional developer?
The second thing that's going to be important is to create a community. Most likely your citizen developers will come from different teams and departments, and they may have never even met each other before. So once they learn these new skills and they're united by your Automation COE’s vision, you need to create a community for them that provides access to ongoing training, support, and they have a way to exchange lessons learned and new ideas. What we also found to be very important is the incentives. You need to celebrate and socialize internally the work and impact that the citizen developers are driving across the business, and you need to recognize them in team meetings and all-hands and show their work to inspire others to join the program.
If you connect those outputs in their work to their performance reviews, or even better, if you reward them with bonuses or prizes, that will further reinforce the continued engagement with the program and that is very important, especially during the early stages when you want the program to stick and scale.
Nischay: Wonderful, Max. It's great to learn about the success of your citizen developer program at Dentsu. So, Max, we've spoken about the immense traction on the demand-side and we’ve spoken about how the Fortune 250 enterprises are rapidly investing in Automation now. Given the traction on the demand-side, there is immense action and crowding on the supply-side as well.
Today we are tracking more than 1,400 platform vendors who are really vying for a place in Automation. In addition, there are multiple types of players as you are already aware, right from the incumbent Big 3 vendors to the technology majors, Low Code/No Code platforms, and even specialist platforms aligned to different technology segments. Given this confusion, how would you advise an enterprise to navigate this complex labyrinth and make a platform decision today, Max?
Max: Yeah, in terms of navigating them all, I recommend dedicating resources to conducting ongoing assessment of how these emerging technologies fit into your Automation stack and support your strategic plan and vision. I believe that the POCs and experimentation is really the only way to truly gauge the capabilities of those emerging platforms. I'm not suggesting that you need to go through the due diligence process with all 1,300 platforms, but at least the ones that are recommended and have already been vetted by industry experts. I would also look at the emerging players on regular basis. AI and Automation market is still an open playing field and everyone has an equal chance to compete for the leadership position. So, you wouldn't want to overlook a platform that might surprise everyone in a few months and offer a new capability that you don't have, but your competitor has already taken full advantage of.
With this in mind, we went through a lot of the discovery and experimentation early on in our journey, because we made it a point to interview and demo three or four new emerging platforms every single week. But a lot of companies today want to cash in on the AI hype. I saw a study recently from a European VC firm which found that 40% of AI start-ups claiming to have AI actually don't use any AI at all. So, you really need to be careful about what is AI, how the company is using it, and do your due diligence.
Nischay: Makes sense, Max, thanks for sharing this. And in addition, I think the space has also seen a lot of M&A activity. There are almost one or two acquisitions happening within Automation on a weekly basis. We have tracked more than a 100 deals which have happened in the last three years alone aligned to Automation. More consolidation is expected to happen in the future as well. Since you've been a part of this dynamic space since long and you enjoy a certain vantage point, what, as per you, are the category of platforms that you feel would win the battle looking at the next five years?
Max: Those platforms that will make AI very accessible to everyone in the business. I mean, democratization of AI and citizen development will definitely rule them all for the next few years. I wish some of these Low Code/No Code platforms were actually around three-four years ago as this is the best way to get started with automation projects today. First of all, they don't require a lot of investment. They're quick and easy to implement. They can work in any business function. They create a short-path to demonstrating value in capturing return on investment and they can quickly scale. But in addition to that, you still need to focus on the top-down use cases as well. So, you need to have a hybrid strategy and tools that support micro and macro needs.
Nischay: Great! Astute observations, Max. So basically, what you're saying is that platforms which have a lot of AI and Low Code/No Code capabilities are the ones which might win in the long run. And I think capabilities, such as being quick and easy to implement, being able to work in any business function, and those that are good at demonstrating ROI and quick to scale...These are some of the capabilities which would be required in a platform to win in the long run.
Max: Yeah. Perfect.
Nischay: And Max, I also wanted to touch upon another important area which is emerging within Automation, that’s around sustainability. So, sustainability has been a key focus area for enterprises and one of the key trends being explored in market around the role of Automation in sustainability. While some enterprises are deploying intelligent technologies to reduce the dependence on paper-based processes, what role do you think Automation can play in moving the needle for sustainability?
Max: Well, Automation has been transformational in areas of economic growth and industry innovation, but it shouldn't stop there. Automation can also have a tremendous impact on society and the environment addressing a wider coverage of goals. With the emergence of AI and RPA, companies now have a choice to either outsource, offshore, or automate the repetitive and routine operations, but companies also have the opportunity and responsibility to improve the lives of those who might get adversely impacted by these choices.
So as an example of what I'm talking about, while I was the Chief Automation Officer at Dentsu we partnered with a company called Autonomy.work which specializes in creating job opportunities for adults on the autism spectrum. Many people on the autism spectrum excel at tasks that require higher aptitude for repetition, pattern recognition, and attention to detail.
So, to help Autonomy.work prepare for that future of work and have Automation technology at their side, we partnered with UiPath to upskill people on the autism spectrum so they can build their own digital assistants. Enhancing their skills with RPA technology presented the opportunity for significant improvements in quality and productivity of their work. But this initiative also provided a great opportunity for Autonomy.work and companies like them to offer new capabilities and create more jobs for people with autism.
Nischay: It's truly fascinating, Max, how automation can actually work towards creating a better society and for the good of mankind. So, thanks for sharing this example with us. Final question, Max, given the action and traction that the Automation industry is currently enjoying, what other top three most exciting trends that you forecast over the next five years timeframe?
Max: Well first, I hope that by 2025 we see a better world where AI and Automation contribute to the well-being of people and there is no more fear of AI displacing people or destroying jobs. Essentially, my hope is that the collaboration between humans and AI-powered systems in the workplace becomes a norm in five years time. As such, I'm most excited about the recent rise of initiatives aimed at democratization of AI and Automation technologies, and that the Low Code/No Code platforms continue to evolve, becoming more intelligent and simpler to use. Without a doubt these technologies will continue to transform the workplaces in the world. So, I hope to see more companies take on up-skilling and re-skilling initiatives to empower employees everywhere to create their own virtual assistance.
And finally, I hope to see quantum computers make it out of the research labs and get us closer to the strong AI, opening doors for new possibilities at new unimaginable scale.
Nischay: Wonderful. So, this was truly illuminating, Max. I’m sure our listeners found this episode as insightful as we did. Right from your journey in the technology industry, to the enterprise focus on Automation, and touching upon what the future holds for Automation, I think your perspectives were truly informative. Thank you so much for your perspectives.
Max: Thank you again for having me as your guest.
Nischay: Thank you everyone for tuning in to this episode of the Zinnov podcast's Hyper Intelligent Automation series. We will be back with another episode with perspectives from another global leader very soon. Until then, stay safe and take care.
Catch Marilyn Krichko discuss all things automation technology - including her journey in the world of automation - which all started with her looking to make her own desk job simpler. Marilyn also shares with the audience her top 10 advice she has for anybody who has just started out on their automation journey.