The ground-shifting digital transformations underway at many enterprises are beginning to transform the executive C-suite as well. Traditional IT leadership roles like CIO & CTO are being joined by new executive titles like Chief Automation Officer (CAO). The elevation of this function to the C-suite parallels the rise in strategic importance of automation, AI, and machine learning to an organization’s future competitiveness.
To better understand what a Chief Automation Officer does and how they can positively impact business outcomes for global enterprises, we turn to Max Cheprasov of Dentsu Aegis Network. As Chief Automation Officer of the world’s 5th largest digital marketing agency, Max focuses on leveraging automation & AI as a force multiplier for Dentsu’s 47,000 employees. He shares with us results from some of the high profile use cases his team has implemented, the surprising way in which enterprises might reap the lion’s share of automation’s benefits, and who the most important strategic partner is for anyone leading an AI and automation practice.
Guy Nadivi: Welcome everyone. My name is Guy Nadivi and I’m the host of Intelligent Automation Radio. Our guest today on Intelligent Automation Radio is Max Cheprasov, the Chief Automation Officer of Dentsu Aegis Network, a 47,000-person multinational media and digital marketing communications company headquartered in London. Max is responsible for leading the company’s digital transformation efforts using AI and intelligent automation solutions to improve operations, processes, collaboration, productivity, efficiencies, and profitability. And we’re eager to dig into his expertise on those topics.
Guy Nadivi: Max, welcome to Intelligent Automation Radio.
Max Cheprasov: Yes, thank you. Thanks for hosting me.
Guy Nadivi: Max, you work at Dentsu Aegis network, which is a sizeable firm, but for those in the audience who may not be familiar with Dentsu, can you please give a brief overview of who the company is and what it does?
Max Cheprasov: Sure. So Dentsu is a global business that is focused on providing best in class expertise and capabilities in media, data-driven digital, and creative communication services. Our global portfolio of agencies includes powerful brands such as Carat, iProspect, Isobar, Merkle, MKTG, and McGarryBowen. And we have offices in 145 countries, servicing 11,000 clients, including 89 of the world’s top 100 advertisers. We’re the 5th largest agency group in the world. And a couple of things to keep in mind that makes Dentsu unique as it relates to our conversation today about AI and intelligence automation is that since 2013, Dentsu acquired over 150 businesses, and we have grown from 15,000 people to 47,000 people. We basically more than tripled in size in five years. And that kind of growth is both exciting and comes with certain challenges if you want to drive change and automation at scale.
Max Cheprasov: So each business that we acquired, of course, came with its own set of strengths and legacy solutions, systems, tools, processes, and best practices, which differentiated them in the marketplace and attracted us to acquire them in the first place. But that kind of process and system fragmentation is typically the main barrier to the adoption of intelligent automation. Of course, what made things easier for me and my team is Dentsu’s strength of integration and the diverse talent and highly entrepreneurial culture. Also given the size and significance of our business, we have access to innovative and cutting edge solutions powered by top technology companies like Google, Microsoft, Facebook, Amazon, and Tencent, and it’s always great to sit down with their product teams and get early access preview to their upcoming products and services.
Guy Nadivi: Your title is Chief Automation Officer, which is a title, I think, we’re likely to see at more and more companies as automation continues becoming more crucial to enterprise operations. Why should organizations have a Chief Automation Officer, and what are the most important benefits one provides?
Max Cheprasov: Yeah, I definitely hope to see more and more companies that employ Chief Automation Officers. But to me, it’s almost like asking 140 years ago when Thomas Edison began commercializing his incandescent light bulb whether you need to hire an electrician to wire your home and office or continue burning candles, right? It was a clear choice back then, and it should be a no brainer decision today. AI’s the new electricity, and think of the Chief Automation Officer as your master electrician. Every business needs to have an AI and intelligent automation strategy and plan to rewire the business and prepare it for the future.
Max Cheprasov: And I think the role of the CAO will continue to evolve as AI and intelligent automation industry continues to exponentially get more and more sophisticated. I think it’s unrealistic for anyone in the C-suite to just take all of that on as an additional responsibility. It really needs to be someone’s sole mission. And I often get asked about my role and who in the C-suite I report to and whether that makes a difference. And in my own experience, I have now reported to a Chief Executive Officer, a Chief Operating Officer, and Chief Technology Officer, sometimes with a dual reporting line. Well, the organization chart in a C-suite may change from time to time, my team’s focus and objectives have remained unchanged. Our ultimate goal is to bring together operational excellence, AI, and automation by closely working together with all of the business functions, regardless of the reporting lines.
Max Cheprasov: So appointing an expert who can navigate across the business and holistically weave AI and automation into every corner of the enterprise is necessary to remain competitive. So every business needs to think about either evolving the role of the Chief Operating Officer or Chief Digital Officer to expand their area of focus to include AI and automation, or if you’re a complex organization and don’t want to rock the C-suite boat too much, hire a Chief Automation Officer.
Max Cheprasov: I think we reached a point in the evolution of intelligent automation when you can no longer delay this decision. There was a new study released by Deloitte a couple of weeks ago saying that the number of companies applying AI has doubled this year, and it also states that 2020 will be a breakout year for scaling intelligent automation. So if you and I were having the same discussion a year ago, I would say that you still have a few months left for research, experimentation, and to tinker with different technologies. However, today you need to have a solid strategy and plan how you will begin to apply AI and intelligent automation technologies at scale, and you really need the Chief Automation Officer to lead the organization on that mission.
Guy Nadivi: Max you implemented an automation center of excellence at Dentsu Aegis. Why should organizations deploying automation implement a COE, and what has been its impact on your automation efforts?
Max Cheprasov: So at Dentsu, our automation COE is a simple mission to unleash the full human potential and take the robot out of the human. It sounds simple, but really not easy. And what we’re trying to accomplish is we essentially want to make teams highly efficient and productive through the use of AI and intelligent automation, and we want to remove the necessary but highly manual repetitive routine low value activities from their day to day grunt work, and give the time back to them to handle more critical and strategic activities that require more creativity and thinking.
Max Cheprasov: And whenever I begin talking about intelligent automation at Dentsu to a new group of people, I start with a slide that has my favorite quote from Bill Gates. He once said, “I choose a lazy person to do a difficult job because a lazy person will find an easier way to do it.” If you think about it, and there is actually a scientific fact and you can research online about it, but humans are inherently lazy. We naturally look for ways to conserve energy and remove complexity from our lives, and we want to reduce the time that we spend on less important manual, repetitive, and tedious activities. So our brains are much better wired for collective social intelligence, innovation, complex problem solving, and creative thinking.
Max Cheprasov: So it’s no surprise that we continuously see inventions that free us up from the routine like the fully automated assembly lines in factories. I mean how many people in the past do you think truly enjoyed the manual process of picking up a part, drilling a hole in it, giving it to the next person in line to inspect, and then repeating that same task a thousand times a day, five days a week, 40 years of their career? I bet not many people enjoy that kind of career.
Max Cheprasov: So the same type of need for automation exists in the office environment. And of course for the automation of the knowledge worker, we’re not talking about the mechanical robots, we’re talking about the software robots, the chatbots, the virtual assistants, cognitive machines, or we simply refer to them as the digital workforce.
Max Cheprasov: So the other thing I’d like to say is that it’s not really a competition, and it’s not about replacing people with robots. It’s how do we take advantage of the latest technology that’s available to us and have the robot and the human work together in a collaboration loop? The main question is how do we build an exoskeleton for the business that effectively combines humans and the digital workforce?
Max Cheprasov: In general, I’ve seen companies take two paths here. You either hire consultants to begin your automation journey, or you try to do it in-house as a COE. And we can debate the pros and cons of each choice, but it’s important to remember that there is no silver bullet here and everyone’s situation is unique. I suggest that if you’re still in the early stages and unsure where to begin, invite some of the top RPA vendors to talk to you about their services and their observations on what has worked and hasn’t worked in companies like yours. Because they now have a ton of experience and can give you valuable and substantiated advice. And do the same with some of the top consultants in the space. You can also attend AI and intelligent automation conferences and try to interview some of the experts in this field. But it’s really not uncommon to spend the first three months, maybe longer over your intelligent automation journey, on research alone. I think it’s a very important first step.
Max Cheprasov: So whether you start to internally build up your own COE, or begin with consultants and then take it in house, don’t rush, and remember that intelligent automation includes a combination of multiple methodologies, and you’re not restricted to just one technology partner. Consider them all.
Guy Nadivi: To your point about lazy people. I can personally attest that some of the most efficient individuals I’ve known in my life have also been some of the laziest people I’ve ever met. So I think there’s quite a lot of merit to that statement. Can you tell us about some interesting intelligent automation use cases your team at Dentsu has deployed and what kinds of efficiencies they delivered?
Max Cheprasov: Yeah, definitely. They are all interesting and exciting. And there is one use case that I’m most particularly proud of because it opened a lot of doors for us at all levels of the organization, and we now refer to this internally as the automation movement.
Max Cheprasov: Very early in our automation journey we intentionally picked a pain point that was sitting in the middle office in a process that tied together our back office functions in finance and our foreign office operations in media. So it was something that most executives in our business could easily understand and relate to. All we had to do was build a robot and illustrate the before and after. And the easiest way to tell a story, especially to the executives, in my opinion, is through visualization. So that’s exactly what we did. We recorded the side by side video of a human worker and a bot performing the same job function. Only the bot was able to achieve exactly the same output, 110 times faster, 100% error free, and it only cost us a fraction of the human labor cost.
Max Cheprasov: So that story not only validated our own COE’s approach and the capabilities of our technology partners, but also highlighted the major opportunity for the business, our employees, and our clients. I think we ended up eliminating close to 95% of human involvement in that process, and through that POC alone, we elevated 25,000 human hours to much higher value activities. And that was just at POC. We were just scratching the surface.
Max Cheprasov: Another use case that I think most listeners will be able to relate to is the RFP process. We get hundreds, if not more, RFPs every single month, and drafting an initial response takes some time and effort. So we built a robot that extracts information from incoming RFPs, applies natural language processing and fuzzy matching to review and classify the information in that RFP, and then extracts from the knowledge repository the best answers for each of the questions in the RFP. The bot then presents a couple of versions, a couple of options per answer to each question to the human expert to select from. And through supervised learning and feedback from the human expert, the bot updates its algorithm, updates its library, and gets better at answering similar questions over time.
Max Cheprasov: But that’s just the start. I think the best part about intelligent automation is trying to compress the time and steps that’s taken from step one to step n in the process, versus just focusing on one of the steps. So as the next evolution for this robot, imagine that the bot doesn’t find the right answer in the library. Who does it ask for for help? How does it find the right expert in the 47,000 person organization to help it answer the new question that it hasn’t encountered before? So our next step is to train that robot to access our HR IS system to find the right expert based on relevant skills and past experience, and collect that response from them.
Max Cheprasov: And there’s so much more that that can be automated after this as well. For example, once the RFP response is fully assembled, or an initial version of it, we can evolve the robot to engage with another AI assistant that schedules a meeting to review the response with all of the relevant stakeholders. And a third AI robot can join the conference call to listen in, transcribe the conversation in the form of meeting notes, upload that automatically to your CRM platform, and then schedule action items in your favorite project management solution. So this isn’t a hypothetical application, it’s what you can do today by orchestrating work and connecting different AI technologies together like building blocks.
Guy Nadivi: Those are intriguing use cases. Max, I’ve read you stated that strong business process management and master data management are the foundations for deploying artificial intelligence and machine learning solutions. Can you please elaborate a bit on what you mean by that exactly?
Max Cheprasov: Right. I believe that a successful AI and automation journey starts with lean business process management. And data is like fuel for the AI engine, right? It’s like oil, it needs to flow through a refine process first to remove any imperfections and waste from it. You need to have a solid plan for how you capture, refine, and manage your data, because every process in your business generates new data. So how you organize and validate that data is very important. You need to ensure that the processes are properly engineered and the data that’s used as inputs is accurate. Otherwise it’s garbage in and garbage out, in which case your machine learning models will be ineffective. If the training data set a solid, you can use machine learning to find anomalies in that data, identify new patterns, make inferences, and essentially make better predictions that will help accelerate the decision making parts of your automation.
Max Cheprasov: I have a hypothesis that one day in the near future, bots will be able to independently operate and make many of our day to day decisions with little to no human intervention in a single enterprise environment that has one rich data set behind it across all business functions without a need for independent and disconnected CRM, ERP, and HRS platforms. If you think about it, all those platforms, whether on premise or in the cloud, all they provide to humans is a user interface to perform certain actions with that data through clicks and mouse movements. And those actions result in either new data created or existing data transformed.
Max Cheprasov: If we’re building bots today that are replacing a need for humans to perform those same operations within those platforms, why would we still need to have access to that user interface? We’re in the early stages of this, but I see us transitioning to a kind of environment where humans work with bots and interact with them and the data using just voice. I mean, we speak three times faster than we type, and keyboards may become a thing of the past or will have very limited use in a couple of decades. Certainly AR and VR technologies will help us get there too.
Guy Nadivi: There’s certainly a lot of people today using conversational AI for a huge portion of their interactions. I see that even happening possibly sooner. Max, in a previous interview you did, I read you say that, “Automation is a catalyst for change, driver of creativity, and procurer of productivity.” That’s very interesting. What are some ways you’ve seen automation driving creativity in addition to productivity?
Max Cheprasov: Oh, of course, if you just automate your existing processes as is and unchanged from the way they’re designed today, you can expect at least two things, a reduction in cost and time that it takes to perform the same job, and an increase in human capacity to perform higher value activities. However, I think a better approach to solution design for AI and intelligent automation is to first ask the question whether the current way of work, that task or that process or workflow, that you are trying to automate is still going to matter and will remain relevant in two to five years from now in its current shape and form. I mean, will it look the same as it does today, or does it need to be rethought and re-engineered? And if it needs to be redesigned, where do you begin? Do you start with that original place where you identified the first use case, or somewhere upstream or downstream from that?
Max Cheprasov: I mean, quite often the pain point you’re trying to address using AI and intelligent automation may have a root cause that needs to be first addressed at the input or a data level somewhere downstream. So find that origin and start there, otherwise you risk ending up with a bunch of short term bandaid type of solutions.
Max Cheprasov: I say that AI and automation is a catalyst for change management conversations because the hype and excitement around the topic, the opportunity, the expected benefits, all serve as an open invitation to have a round table conversation with the process experts, executive stakeholders, and end users. They all need to be engaged in a discussion of what you’re ultimately trying to achieve with AI, and how you will build a sustainable and durable solution based on the human need behind it.
Max Cheprasov: Don’t be afraid to take a step back to reevaluate your current designs in the process before you automate something and analyze the entire assembly line. Take it apart, step by step, function by function, and engage everyone in that redesign conversation to reassemble it for the future of work with AI and intelligent automation ingrained. This is called design thinking, when you focus on your automation solution design on the people you’re creating it for. Understanding their core needs, brainstorming with them, and prototyping together with them. That leads to better product, services, and internal processes.
Guy Nadivi: Hmm. Given your expertise in this field, what do you think are the top two strategic reasons organizations should automate?
Max Cheprasov: Well, the first reason is simple. In the age of AI, the second place will no longer be good enough. Whoever in your competitive field gets to the winning AI and intelligent automation formula first will leave everyone else so far behind that it will be really expensive to try and catch up. I mean, the front runners will have impressive tools that will enable them to analyze and process more information, they will make better decisions, produce more output. They will do it much faster and much cheaper than those who don’t have that winning AI formula.
Max Cheprasov: And it’s not just corporations that are competing in this race, right? The governments around the world understand that AI is a foundational technology that can significantly boost competitiveness, increase productivity, protect national security, and even help solve societal challenges.
Max Cheprasov: So that’s the first reason. The second reason is that the next generation of employees and the younger workforce in your business have much higher expectations for their work life. I mean, Generation Z is the first fully digital generation, and they expect high tech solutions, not just in the palms of their hands at home, but also in the workplace. So if you’re still doing things on paper or if you try to assign a Gen Z to do repetitive routine, copy and paste type of activities, or any transactional activities that don’t require much creativity or thinking, you won’t be able to keep them around for more than six months at best.
Max Cheprasov: So you need to prepare your business for that, and I think the best way to accomplish that is by engaging millennials and Gen Z in the automation solution design process. If you invite them to participate very, very early in the discovery conversations and excite them about the opportunity to be involved, and just focus on redesigning the workplace around them, that benefits them and ultimately your customers.
Max Cheprasov: And beyond thinking about the incoming workforce, your AI and automation strategy and plan will also need to address how you will manage your current workforce that’s impacted by automation. You know, how will you be addressing the changes in their current job roles brought on by automation? Will you be reducing your human workforce in one place of the business but increasing it in another function? What will that transition plan look like? How will you be up-skilling or re-skilling those impacted by your automation efforts? So those are very important questions. In my opinion, the most important strategic partner for anyone leading AI and automation practice should be the leadership team of HR from day one.
Guy Nadivi: Interesting. When you evaluate a process to be automated, is there a minimum ROI you need to justify automating it?
Max Cheprasov: Yeah, that all depends on what you want to achieve through automation. Is your ultimate goal to reduce costs or improve employee satisfaction and provide a better work life balance? Or is it about reducing risks? Is it about eliminating human error or improving compliance? Most ROI models will be about cost cutting, of course. Some will be about improving work life balance, or quality of output, or customer satisfaction. But not all benefits will be easy to quantify or track and measure.
Max Cheprasov: At Dentsu, we create the business case for every project, and it’s not because we need to justify the investment or predict the financial outcome every time. But it’s really because these projects can take weeks and months to deploy, in some cases longer, and the main purpose of the business case is to capture and to remember the reasons we decided that it’s important to automate in the first place. What is the ultimate destination? What is our why? Why did we decide that it’s important to automate? What alternatives have we considered? What are the risks associated with doing nothing? The answers, even within the same company, within our business, may be different for every project, every time. And our standard ROI model consists of about 20 factors, from the cost of investment perspective going into automation.
Max Cheprasov: We consider everything from internal staff costs, infrastructure costs, security training, third party technology license fees, just to name a few. And then from the benefits side, it’s the labor savings, cost avoidance due to 100% accuracy reduction, and time needed to conduct compliance audits, work life balance, employee satisfaction, customer satisfaction, the new business opportunities that you generate as a result. And I think the ROI model for every business will have many of the same factors, but in every case the weight or significance that you assign to these factors may be different from project to project.
Max Cheprasov: I would also say that, from our experience, if all you focus on is the labor savings and nothing else, you can expect to break even on average in six to 12 months depending on the complexity of the business case and the experience and maturity level of your automation team. However, the lion’s share of the benefits comes in a form that is hard to measure or quantify sometimes. For example, how do you measure the elevated potential of your human workforce, the creativity boost that you just gave them, or a moment in time when someone actually had the time to apply critical thinking to avoid a major crisis?
Guy Nadivi: You just spoke of work/life benefits, which makes me curious. What do you see as the future of work at Dentsu Aegis Network once AI, machine learning, bots, and other automation tools become more common? What’s a typical day in the life of a Dentsu employee going to be like five years from now?
Max Cheprasov: Oh, I see a beautiful future, and I have a difficult time every day trying to balance fantasy and reality. I think by end of next year, speaking realistically, we’ll start to realize benefits of AI enabled platforms and tools at scale. And we’ll see that intelligent automation will start to become a way of life as we played with the idea of introducing a citizen developer concept, where employees can create their own micro automation solutions in a low code to no code environment, while the COE focuses on top down opportunities and larger scale opportunities.
Max Cheprasov: I think over time, intelligent automation solutions will become more and more intelligent and adaptive to the processes being automated. So I see us becoming an environment with self-viewing, self-healing, and auto-optimization methods becoming widespread, and automation inherently becoming smart. And this is what I call an AI-driven enterprise, where AI is in the DNA of the business, and that is where we are aiming to be in five years.
Max Cheprasov: Like I said, I think it’s a beautiful future where every employee has a virtual assistant that is smart enough to pass the Turing test, and humans working together with cognitive bots becomes a new norm. And this is not a fantasy anymore, it’s already happening. It’s just happening in very few places and in very controlled environments. But as our AI and intelligent automation partners continue to evolve their products and solutions, we know that kind of future is realistic.
Guy Nadivi: I think if you had drawn that vision 10 years ago, it would’ve still been considered more science fiction than fact. But of course as we’ve come to learn, there is no such thing as science fiction, there’s only science. So I think your vision is pretty realistic at this point.
Max Cheprasov:Yep, true.
Guy Nadivi: Max, for the CIOs, CTOs, and other IT executives listening in, what is the one big must have piece of advice you’d like them to take away from our discussion with regards to deploying automation and AI at their organization?
Max Cheprasov: Well, I will kind of repeat what I said in the beginning when I said that AI is the new electricity. Every business needs to have an AI and intelligent automation strategy and plan to rewire the business and prepare it for the future. Do you have one? If not, you need to have one soon if you want to remain competitive. And I think the biggest challenge for all of us today, of course, is the shortage of experienced professionals in this field. But very soon it’s going to be extremely hard to keep up with all of the emerging AI platforms as more and more VC money is being invested in artificial intelligence startups.
Max Cheprasov: So no matter where you are in your journey, I believe that research and experimentation with these new technologies needs to remain a top priority and focus. And in turn, that should keep your AI and intelligence automation strategy informed and constantly evolving. It’s not going to be a static plan. So be prepared for a bumpy ride. An exciting one, too.
Guy Nadivi: Good advice from somebody with deep expertise on the subject. All right. Looks like that’s all the time we have for on this episode of Intelligent Automation Radio. Max, thank you so much for joining us today and being the first, but hopefully not the last, Chief Automation Officer we’ve interviewed. It’s been great having you on.
Max Cheprasov: Guy, thanks for inviting me. It was a pleasure.
Guy Nadivi: Max Cheprasov, Chief Automation Officer of Dentsu Aegis Network.
Guy Nadivi: Thank you for listening, everyone. And remember, don’t hesitate, automate.
Chief Automation Officer at Dentsu Aegis Network.
Max Cheprasov is the Chief Automation Officer at Dentsu Aegis Network, a multinational media and digital marketing communications company with 47,000+ employees in 145 countries. Max has 25+ years of experience within the Digital Economy, specializing in digital transformation, operational excellence, and AI-powered automation. At Dentsu, Max founded the Automation Center of Excellence (COE) that develops and deploys highly effective machine learning systems, narrow AI and intelligent automation solutions (IPA / RPA), augmenting and enhancing employee experience with smart digital assistants, and improving efficiency and effectiveness across the business.
Max can be found at: