Episode 7 – Elevate Human Potential With Innovative Automation
Feb 04, 9:00:00 am
FortressIQ | Intelligent Insights for the Modern Enterprise
Episode 7 - Elevate Human Potential With Innovative Automation
Today’s guest is Ben Nabulsi, a process data scientist at Dentsu. Host Jon Knisley, long-time technologist helping companies win the market with emerging AI technologies, talks with Ben about how we can elevate human potential with innovative automation. They answer your questions about the integration of this powerful new technology with companies and how it can help them reach their corporate goals and greatest potencial.
Dentsu is one of the largest global advertising firms in the world. Because of Ben’s position as a process data scientist, he is able to sit in a front row seat as Dentsu’s innovative automation has had a positive effect on the workforce. Ben agrees that developing an AI and automotive strategy is essential to the success of any modern business.
Every company, no matter what size, knows that in order to remain competitive, you have to embrace change. Most transformation programs in the last decade have focused on customer experience, but today, indicators suggest the emphasis moving forward will shift toward back-office operational excellence. FortressIQ customer dentsu is a company that thrives on innovation and an entrepreneurial spirit, which powers its Automation Center of Excellence and that team’s mission to elevate human potential across the organization.
A day in the life of a data process scientist
How dentsu and Ben were able to utilize FortressIQ in analyzing their data
How the “new normal” has affected the transformation programs
How to build and maintain trust with stakeholders when it comes to the new technology
Credentials, security, and other types of measures to ensure security and privacy of the stakeholders
The “people element”
The role of business analysts in this platform
Expansion and the future of the program
Advice for companies who want to follow in the same path as dentsu
If you enjoyed this episode, subscribe and check out our series at fortressiq.com/podcast. Thanks for joining us today on hello, Human.
Full Episode Transcript:
Given our guest, this promises to be a very interesting and lively discussion. I can ensure you that you will leave with at least one nugget of insight to impress your colleagues as well as help make your organization more resilient and competitive in this new environment that we are all learning to navigate. With me today is a true industry leader Ben Nabulsi who is a Process Data Scientist at Dentsu. One of the largest global advertising firms in the world with sixty thousand plus professionals spread across more than 145 countries. Often it’s a cliche, but in this case I can promise you it’s true. The agency really thrives on innovation and entrepreneurial spirit, which extends to its Automation Center of Excellence where Ben works. The team’s mission there, to elevate human potential with innovative automation across the enterprise, that’s our topic today.
I’ve provided two Bitly URLs, the first one—FIQ-Dentsu—that links to the article in Business Chief Magazine. That was the impetus for this session and detailed the work and success of Dentsu’s Automation Center of Excellence, which we will hear about today. The second one—FIQ-HFS—that links to the HFS research report that recently main FortressIQ as a top three process intelligence platform. Which is one of the key components of Dentsus technology stack which powers their COE.
One final housekeeping comment before we jump into the discussion, Ben and I want to make sure this session is helpful for you. Submit any questions or try to make comments to us anytime and we’ll do our best to work them into the conversation. This is going to be a purely conversational format today, so we’re not going to bore you with any slideware. We’re just going to jump right into the conversation/discussion.
Thanks again for joining us today and Ben welcome to the program. You’ve got this cool title—Process Data Scientist—really combining two hot trends in business and IT. I did an admittedly non scientific search on LinkedIn and uncovered just a handful of people with that same title. I’m sure we’ll see many more in the coming years. But what that day in life looks like for a process data scientist? Can you give us some background on your role? How did you get into it? What do you enjoy about the role? Where do you wish you could spend some more time?
Hi,welcome to hello, Human, a podcast to explore ideas and feature humans working in AI and technology.
Jon: Ben Nabulsi, Process Data Scientist at Dentsu joins us today on the hello, Human podcast where we discuss the latest topics in artificial intelligence and how it’s being applied in the real world. I’m Jon Knisley, the host of hello, Human and a long time technologist helping companies use next generation digital solutions to win in the market.
A big thanks to FortressIQ for sponsoring the program and be sure to subscribe wherever you listen to podcasts. In this episode, we are going to explore elevating human potential with innovative automation. Dentsu is one of the largest global advertising firms in the world with 60,000 plus professionals brought across more than 145 countries. Often it’s a cliche but in this case I can promise you it’s true—the agency thrives on innovation and in entrepreneurial spirit. That extends to its Automation Center of Excellence where Ben works.
As the process data scientist, he’s had a front row seat and provides incredible perspective on the positive opportunities and changes in the workforce as a result of innovative automation. He really understands why developing an AI and automation strategy is essential to the success of every modern enterprise. Welcome to the program, Ben. Thanks again for joining us on hello, Human. You’ve got a pretty cool title and I’m a bit jealous—Process Data Scientist.
Combining two of the hot trends in business in IT today, I did an admittedly non scientific search of LinkedIn and uncovered just a handful of people with the same title. I’m sure we’ll see many more in the coming years, but what’s the day in the life look like for a process data scientist? Give us some background on your role. How did you get into it? What do you enjoy about the role? What do you wish you could spend more time doing?
Ben: Hey Jon, thank you for having me, I’m excited to talk to you today. My name is Ben again. I joined Dentsu more than five years ago now. I’m not big on titles, but the nice thing about Dentsu is we have specialization and whenever I meet someone to be really into a specific role, they work to get that role to serve the business. That’s been the start of how we, from a data structure background that I had to being more into the process of data science and getting into the process.
My daily journey starts with looking at the recent mining, that’s from the night before that we have. See if I need to refine the mine or need to move into the documentation process. It depends where we are in the stage of the cycle. If we need to refine the process, we’re looking at the database and just looking at the signals we have, to refine it and to mine more. It’s always been into that mine, finding the process. It’s like hunting for the good process, high quality process, so we can document and share with the business, and get the feedback. There’s a lot of other things too, but there’s more into the process and getting it to that level.
What I’d really liked to spend more time on was research. That’s one of the things I really enjoy doing—researching. Everyday there’s a new technology machine learning model there and how we can use the data that we have to see how it reacts with it. Or how can we really look at the data that we already have in different lenses. I think that’s one of the exciting things about having such a broad data set.
Jon: I think that’s really cool. You sort of blend the art and the science together in that role. It’s not purely structured, you think of data science and being so structured, you’ve got this pipeline that you follow. But you’re able to bring in that sort of flexibility and that art to the process as well.
As I mentioned in the opening, the real trigger for this conversation was that recent article in Business Chief Magazine and really did a great job of highlighting Dentsu’s Automation Center of Excellence, where you are. I think that the stat that really jumped out at me was, your team was able to map, model, and document. I think it was 2200 processes in less than 5 months. The amazing thing is with just two people using the FortressIQ platform. Really incredible output and I’m sure it’s been a journey to get there. Simple question, no real easy answer—how’d you guys do it?
Ben: It’s something everyone asks about. It’s definitely a combination of deep engagement that we’ve been going through. Whether it was with the FortressIQ technology and the team from different disciplines to the business side of Dentsu—whether it was the executive boards, how close we work with, and the leadership of the agencies—because we run it for multiple agencies. And then having our own techniques from our understanding of our applications and the user’s workflow. For having techniques to target these processes, all of these elements help us to target these processes like never before.
Jon: Kind of exploring today’s environment a bit. We’ve heard many companies really experience this accelerated transformation during the early stages of pandemic. I think at one point I saw Microsoft reported that they had two years of digital transformation in two months for instance. Obviously Dentsu’s transformation was well underway prior to 2020. How has this new normal impacted the transformation programs? I would assume there’s some extra interest from Dentsu leadership these days. Can you provide any stories from the front lines?
Ben: Yes, so we are lucky in a couple of senses, we were not planning because no one knew about the work from home change that’s going to happen all at once. But we’re running this program from 2019. We were already familiar with the systems, tools, and we have our own process going. When work from home happened in March, we’ve definitely seen a shift in how people interact or how people are using their applications.
We start to do, okay, let’s do something like work from home analysis. How did it really impact us as a business? Whether it was through the specialized or the productivity apps. That was one of the things we did, because we already had an observation going where we’re having a team looking into a team, and they had to go work from home—and we all did. That was very insightful to see how their behavior had changed.
Also it helps with this program specifically because sometimes you send out a link without context for your colleague when you need to have a conversation into this. That has been taken away from working from home. So they have to really put the context, act like it’s talking to others. That’s helped us understand more the context than before when they do certain things. I’ve seen it from our side from the system and the work we capture. These were one of the interesting things we faced and we’re looking forward once we go back to the offices, hopefully next year or so, how the change in behavior will happen? How did it impact on a short or long-term the application usage of how people behave?
Jon: I’m just curious on that note. Have you guys set a return to office, any timelines in place yet? You mentioned sort of next year. Did they say anything specific yet to attend to? Or that’s still to be determined at this point?
Ben: To be determined. I don’t think anyone knows at this point. We always hope to go back to normal at some point, and we’re hopeful for next year. But there’s nothing specific, because no one knows.
Jon: Traditionally process optimization, automation is seen as fitting more traditional industries with very structured data and procedures. Obviously media and advertising by its nature, is much more agile and creative. I think oftentimes many people are surprised to hear an agency is a real leader in this field. What do you attribute the key success factors to Dentsu’s transformation success?
Ben: There are multiple things. I will go back to the main points that I’ve mentioned about the business engagements and the technology, but I will start with the technology because if you look at the industry of process mining you have solutions that are trying to capture information from a one system, like salesforce or other systems that are unified. Our industry is not unified in one solution. That’s one of the uniqueness of Dentsu. Working with different clients that we have, we have to use different systems to support our clients and their uniqueness.
Now with the solution of FortressIQ, we were able to tap into the user logs, not just the system logs. Which is a different way of looking at data. It’s not connected. When you look at the system logs it’s usually logs and diagnostics of the user experience there. Now we’re looking at the behavior of the user with application on their uniqueness. That was one of the success factors, the technology is unique. We have the business engagement that we had when I mentioned before, we have the SLPs always in place for, we know it’s there but not everyone will probably use them or follow them specifically.
But for the first time we’re able to see the process of what it is, not how it should be. That’s the realistic take of the actual process. All of these help when we transform or try to transform the conversation are more real, are more based on numbers, based on common behavior of practice, not just how the system menu or the manual is.
These kinds of engagement with the business, I will also talk more about it. It’s very unique because of the support and investing in the future kind of situation where you try to establish this database to serve current needs. But we always come back and remind and really look at the data in different ways, depending on the team and depending on the problem you’re trying to solve.
Jon: Kind of along those lines. Obviously the perks of your platform, we can move it around to different systems and different applications. There’s no real integration we need to do on the backend because it’s essentially recording the activity that’s on the screen. Do you guys find that helpful in your use of the platform? Are you looking often at multiple systems across different users, or are you kind of limited in terms of which applications you guys observe?
Ben: There’s no limit, that’s the thing. We’re seeing the actual interactions. Every time we do a cycle, we discover a new application that we never knew that it’s being used, just like it happens and that’s realistic. The teams have to use these solutions to support their work and we get that high level of events that are using it. Then we get curious to understand, why is it unique? What’s different from the main one that we have? Do we need to develop support for the current system to make it takeover these small functions? It’s kind of good feedback for what we are used to and what reality is. Does it make sense?
Jon: Yeah, that’s great. Let’s turn to the human arm a little bit. There are a couple questions that came in on this topic as well so we can sort of bounce off those as well. Whenever there’s talk about process discovery, one of the first questions that get asked is about privacy and security. It’s fairly obvious and easy to understand and comprehend why. When you have the traditional manual approach of workshops and time motion studies, some of these interviewing you or watching over your shoulder. The participant decides what to share with our more modern technology based solutions out there, the trust issue becomes much more critical. How do you build and maintain trust with your stakeholders that you’re observing?
Ben: It’s a good question, it’s one of the trickiest things and this kind of culture change, because everyone is like, oh, you’re looking at my monitor, are you looking at my screen, you’re seeing what I’m doing. But it all starts with explaining the technology and it could be as simplified as much as possible. But always be ready if someone challenges you with how you’re dealing with this situation? Having this very honest conversation we do, that’s every group we start with. We have a kickoff and that’s something from this patent case, I work with Michael Stockton and basically talk to the team. Any questions they have, they can ask that’s on the spot, or they can iMeet separately and we can answer these questions and concerns. Anytime they feel uncomfortable they can say, can we be off, and we’re always open, they participate into this program, we don’t force on anyone.
Other things we also do is take extra steps. The way we look at it in different ways is we don’t want a lot of noise, we’re really into the process, we’re really into the applications that the company invested in and to lose these functions. We took extra steps to remove any events that are not relevant. We don’t want to see anything private. Even though it’s a company machine and we all use it in different ways because that’s what we have. But we took that step, where the user can submit anonymously a URL so we avoid it or we can’t target specific applications. There’s a lot of ways we can implement this program to choose being with that compliance.
So these extra steps along with managing the deployment itself. I think that’s one of the key things we took on. As we don’t go through our teachers, but from the software on scale. We have more than 15 users, we implemented ourselves. So we make sure that everyone is on-boarded and if one is off-boarded. If anyone’s having any issues, they can directly reach out. There’s no wait for the ticket from the RD to come back. We took that ownership and it helps them to feel confident when you send an email, okay, you’re off we’re done. They feel like, okay, it’s really done.
That’s the build of trust, because the same users you might come back in a couple of months because they participate in different applications or a new application. This helps, we believe. More than 200 people, they’ve participated in this program. It’s been more accepted than before or less concerned because they know our purpose is really the business, not trying to get anyone in trouble.
Jon: Yeah, that insight is great. Actually came up in some conversation I had yesterday with some customers and the idea that this change management culture issue is so critical. But ultimately, at the end of the day, every organization has got to deal with it. It depends on the culture of the organization. But I think once it’s explained to people and that change management program comes into place, we have a very good history with working with organizations and making sure that the stakeholders are comfortable with the technology and the approach.
Along those same lines, there were the questions that came in about certifications and credentials, and then I’ll give you a chance to take a break here and answer this one about what credentials we have in place given this type of observation that’s going on. A couple different ways that we handle security, at different levels we have allowed and denied a list of applications. We also have a gateway, a masking appliance that is a privacy enhanced gateway that we call PEG that allows us to mask any sensitive information on the screen.
On top of those two layers, we also have a number of certifications, regulations, and guidelines that we follow, headed by the GPPR, […], ISO 2001 always working on other ones as well. We take security and privacy very seriously. Today, always comes up in conversations but never has presented a major hurdle for programs to move forward. Ben, we also got a couple questions on this too, so I’m interested to hear your take on it. The people element often gets forgotten in terms of project success and the consultant, they always look at the golden triangle of people, process, technology.
In your role, I imagine that you need to work both top-down and bottoms-up. Can you provide any guidance on collaborating with management and leadership as well as the subject matter experts and process leads in the business units to meet objectives and achieve those targeted goals that you’ve got?
Ben: Yes. That’s one of the fortunate things and luckily, we have the Siri of automation. It’s really unique and that’s becoming more of a trend in the industry, but we have a very unique team. In my team, for example, we have the VP of automation. It’s like Max, and we have and we have Bryan. We all work together and connect with the executive board to understand what they’re trying to do and how we can help and whether it was automation or process link. Having these conversations with the CEO like Lucas Cridland and Sean Power or other executive leadership and having deep engagement with the business lead, like our executive vice president, Michael Stoeckel and Robert Hannan. Having these sponsors and supporters where they understand what we try but we can’t be doing but they see the vision of where we can go year over year, that has been very helpful to support this program and future applications and we always look at what’s next, how we can tackle new challenges or review things.
The human element has been one of those most important things beside the technology. Even working closely with FortressIQ, data science team, or product team and getting that support of understanding at the backend how things work and how we, together, can get the business objective working with the technology and getting to that goal. It’s very unique, it’s very engaging, and it’s been a thing so far. It’s exciting to work with such dynamics.
Jon: You mentioned Max Cheprasov, your colleague at Dentsu and I think it’s interesting to hold that title Chief Automation Officer, really that C-level type title in, again, we already talked about it, sort of a more agile and created industry, you wouldn’t certainly expect that. I think that, again, shows that seriousness that you guys take with this area. Max has also talked about this multi year roadmap that he’s created to really weave automation into every process at Dentsu by 2025. He’s talked about making AI and automation really a part of Dentsu’s DNA which I think is great.
Also, we had a question come in that’s kind of related to this around, Ben, if you could address what you see as the role for business analysts, process, discovery, and documentation moving forward. What efficiencies do you see them gaining using tools like the FortressIQ platform? And challenges as well.
Ben: Yes. It goes back to the uniqueness of this technology. Because usually business analysts or the kinds where you look at the past logs or it’s just the archive of data. Now you’re looking at different data sets, different disciplines. It’s ready and normalized data for you to see and use the mining AI power to drive your queries, that’s very powerful. Having such tools and to drive your analysis or your search, helps you check things, if you’re trying to analyze something, and helps you to see it on scale.
Having 200 people, we have 3 TB of data, that is not an easy thing if you just want to use the legacy applications to look into. It’s a huge database. How can you use AI models and techniques to get to what you’re looking for quickly and accurately? That’s a very powerful tool for anyone interested in analyzing the data.
Jon: Somebody asks us to go back to the static throughout earlier from the article that was the 2 people document and captured 2200 processes in 5 months. The question is, I want to make sure I get it right so I’m going to… How many people process data from the 2200 processes in order to make it in a form of information that becomes usable in decision making? I’ll let you think about that answer for just a second and I’ll preface it by saying the article went on to handle that amount of information and data, capture it, document it all if you were doing it periodically manually. I think the article noted that it would take at least 30 business analysts to get that same level of detail that was captured through to the system using the system and two analysts.
I don’t know if you want to add anything further on to that comment, Ben?
Ben: Yes. In our experience, what we’ve done and that’s working with Michael and the team, is basically all the PDDs we’ve created and PDD stands for Process Definition Documents. We have more than 140 of them. We made them available in a SharePoint location and we’ve been engaging with the business users referring to these PDDs when we’re targeting a process.
Sometimes, even if they don’t really realize these processes or understand them as is, they come back. It’s like a point of reference. Having that library of documentation there, open for anyone interested to see these processes, how they are done, how many people have done it, the average times, that’s very powerful for anyone. We keep it open for anyone to look at it. It’s not restricted to the boards or restricted to a specific business, that’s a knowledge sharing where we’re trying to do was really unique.
Jon: The culture question is driving a lot of interest. We got another question that’s about has Dentsu leveraged mining process intelligence across other markets? If so, what have you seen and what differences have you found in Japan—that’s your headquarters and their culture is very distinct.
Ben: That’s a very good question. Actually, our experience in the US has attracted a lot of interest from different regions. We are currently talking to APAC like Australia, Asia, and UK, Ireland. They are all interested in and we’re working with them to expand this program next year. We had an observation happening for our RT users in the UK. It was a unique experience because we currently been looking at the media side of the transactions.
The Canada RT side was a new experience. It’s different because now we’re able to see the end user, the mid users submitting a ticket for RT and how the RT is doing it. It’s like completing the picture. That’s what I was saying, the database they have now, the more you have different groups and observations, the picture becomes more clear. You can see a loop back who’s handling the process, how it’s been taken. It’s all connected.
We’re looking forward to different users. That’s one of our missions, to include more differentiation. They’re from the regions. The US is good at certain things, other regions are good in other things. How can we learn from each other? We’re using the same systems, why is it different productivity or time? These are future plans we’re looking forward to.
Jon: That’s great. I think you already touched on this next question we were going to jump to. As we start wrapping up, we talked openly about your role in activity as a process data scientist, obviously, the team is doing incredible things in the organization that add significant speed in scale. You just talked a bit about going to different regions, to different areas. If you dust off that crystal ball, where does the program go from here? What else would you like to tackle? There are different news cases to start looking at different systems, obviously. Dentsu goes through many acquisitions given that’s the norm in the industry. Is that an area that you’ve thought about using the technology?
Ben: Yes. I’m very ambitious about how we can drive change. That’s one of the things we always aim for. One of the things for the future is how we look at this program . One of the hardest things to get machine learning and getting smart applications or more predictive applications is getting good data sets that you can chain the models. That’s been a challenge usually in the industry—getting a data set that is normalized and ready to fit a future application.
I think I’m looking forward to the platform itself becoming more, hopefully in the future be more predictive or how can we use our custom models from the data we’ve collected, from the data we normalized and mined for to fit these applications that can help us either predict or look at things in total different dimension. That’s one of the exciting things. This year and last year has been really perfecting capturing the data, understanding the workflows and having our data marks and the places we want.
Next year, I’m hoping we get to how we can confirm, how we can have diagnostics of the process, how we can say we can drop efficiency more if we change X, the efficiency will change. If we can tick the time from doing certain things can allow them to do more transactions, specialized things. Because our goal by the end of the day is giving more time for the specialized user to do the things they like. Not the repetitive things or the things they are not into. Process mining and automation works together to achieve that goal, help the business understand what are the things that maybe they don’t see and how we can drive the small tasks to automation and allow them to do the things they enjoy.
Jon: Dentsu is obviously having significant success around process intelligence. Give me your top three pieces of advice for companies in any industry that want to get where Dentsu is but are a little earlier in the journey.
Ben: The three pieces of advice from our experience, the first thing is try. Try this technology of process mining. Even if you don’t have expectations or you don’t know exactly what your objective is, just try it, test it. The dots will be connected later on even if you’re not really sure of. Try to invest in these kinds of technologies because the dots will be connected. Some users will look at the data and the story will make sense and it will drive you next. Taking the first step usually is the hardest and having this process internally, how can we deal with how we can work with this application, how can we engage people. Because you can’t just flip the switch and just be too quick. You need to engage people and buy in to drive the technology, drive the business, drive the transformation. The first advice would be try not being hesitant or too slow on it.
Second thing will be privacy is concern. We all have this liberty that we want, we’re worried about privacy. Avoid being limited to such applications because we all use equipment of the office to drive the business, to support our work. We’re not targeting specific areas that the business is not interested in. Our goal has been driving the business for efficiency and its goals. Privacy shouldn’t be the limiting force of not having this program going. I think that’s one of the pieces of advice and the tricks we’ve shared.
The last thing is automation and process mining are hand by hand. That one’s becoming clear to the industry. If you don’t have Siri automation, there must be a starve there. Because automation and process mining are hand by hand. The technology can go along with and learn from these experiences. But investing in both, you need it because if you don’t know exactly what you’re trying to automate, it’s all become theoretical and trial and error. But if you have an actual process on the actual people who do it, then the automation would be easier.
These are the three pieces of advice from the very humble experiences from experience that I have.
Jon: That’s great insight and a great point to end on. To recap today’s conversation with Ben Nabulsi, the Process Data Scientist in Dentsu’s Automation Center of Excellence, done correctly innovative automation drives positive opportunities and changes for the modern workforce in developing an AI and automation strategy is essential to the near and long term success of an enterprise. I’m going to put two resources in the show notes. First we’ve got the Business Chief Article on Dentsu’s program, that was the impetus for this session and details the work and success of Ben’s team. The second one is the recent HFS research report that named FortressIQ as a top three process intelligence platform and which is one of the key components of Dentsu’s tech’s deck which powers their COE.
That’s a wrap on today’s show. Thank you Ben for joining me and FortressIQ for sponsoring. I’m Jon Knisley and this has been hello, Human.