TRANSCRIPT
EPISODE 37: Sidney Madison Prescott
Jim Freeze Hi! And welcome. I’m Jim Freeze, and this is The ConversAItion, a podcast airing viewpoints on the impact of artificial intelligence on business and society.
[UPBEAT MUSIC]
Today on The ConversAItion, we’re joined by Sidney Madison Prescott, Spotify’s Global Head of Intelligent Automation. Sidney is an expert in all things AI, automation and RPA. Last year, she even co-authored a book on how to build robots using real-world prototypes.
Now, she’s sharing her wealth of knowledge at Spotify, teaching stakeholders about the benefits of intelligent automation and upskilling Spotify’s workers, while empowering them to identify opportunities for automation. She’s here to tell us all about it.
Sidney, welcome to The ConversAItion!
Sidney Madison Prescott Happy to be here.
Jim Freeze Some may say you have a fairly non-traditional background for somebody who’s an expert in automation. Your experience covers philosophy ethics, political science and business and I understand you’re even pursuing a master’s degree of legal studies at Cornell right now. So can you take us through your background and how it led you to Spotify?
Sidney Madison Prescott Absolutely I definitely would categorize myself as someone who entered into the technology sector from a non-traditional background. I did begin my educational pursuits with the intent to actually go to law school eventually. And so to that end I took a bit of a serendipitous turn in terms of my career path by taking an internship during my junior year of college. So I was a philosophy major, pre-law concentration and I wanted to get some real world knowledge in relation to business and how it relates to general counsel and more specifically contracts. And so I had this incredible opportunity to take on an internship within the financial engineering world, and this was with a financial payments solutions provider. And they were a technology firm, a global firm, that specialized in really understanding the nuances of payments from a retail perspective.
And so initially I was brought in to actually read software contracts and so my goal was to look at the details of the contracts specific to licensing for software, and to understand where we had gaps in terms of utilization. And through that engagement as an intern, I began to really start to dive into the technology sector more specific to everything from big data, to kind of the data transition between those new virtual machines, SaaS systems, and the difference between the SaaS or the Cloud applications versus legacy systems.
And through all of that journey I quickly found a passion that I really had never unlocked before in relation to technology. So as that internship shifted on into my senior year of college, I had a decision to make about where I saw my career going next. And I made the decision, I thought about how often I considered the business problems that we were dealing with within the firm even after I left for the day, and how much the excitement of the work that we were doing really ignited my creativity. And I wanted to continue to pursue that. And I also had great feedback from my mentors and at that point I made that decision to pivot away from a career in law over to a career in technology and that segued into coming on board the firm.
And I led a global data quality and governance program which really gave me a great insight into how to manage the distinctions between disparate systems, ETL, what do we mean when we talk about extracting transforming data. And it also gave me a better understanding of the pain points of business stakeholders in relation to technology initiatives and how to bridge the gap between our business stakeholders and our engineers who sit on the technology side looking to facilitate automations to help the business. So it really gave me a really nuanced way to navigate through and learn the ropes about the distinctions between those two pillars, which are present in most firms, and also really how to think outside of the box in relation to solving business problems through various automations.
Jim Freeze So that is fascinating. I’m going to demonstrate to our listeners how much smarter you are than me because I actually went all the way through law school and graduated and practiced for a little while before I made the pivot to, I obviously work for a conversational artificial intelligence company now. So you’ve figured it out much sooner than I did so good for you.
Sidney Madison Prescott Yes, it was definitely an interesting pivot. It was one I did not see coming, but I am so happy that I listened to myself, to those passions, as I said, that I uncovered during that time. And I also listened to the feedback of industry peers that were privy to my work and who really gave me the encouragement to pursue a career in technology.
Jim Freeze I think that’s fantastic. So, using the word pivot again, let’s pivot it over to your current role. So you oversee Spotify’s intelligent automation strategy, which includes everything from RPA to AI components like machine learning. Quite an undertaking. Can you talk a little bit about your work at Spotify, what that entails and what your team’s goals are?
Sidney Madison Prescott Absolutely so my team sits within the financial engineering function within Spotify and so we are a part of the R&D mission at the firm. And more specifically, the team arose out of a need to look across all of the different business processes that occur on a daily, weekly, monthly basis within Spotify and to see how can we actually lift off the manual, repetitive, tedious work from our Spotifyers? And how can we make a change in the workflows so that they can become less manual and more automated?
And the way that we are taking on that initiative is, as you said, through the use of several different tools. So we’re leveraging robotic process automation, which in essence are software robots that mimic the end user’s ability to navigate through different systems. We also combine the robotic process automation with machine learning which can help us parse data that may or may not be machine readable, so that we can then pass it off to the bots. And then we are now starting to investigate diving into artificial intelligence and this is where we are looking at how can we engage our business stakeholders in processes that require more cognitive decision making? And how can we still automate those use cases rather than pushing them away because we don’t have the capabilities from a tooling perspective?
Jim FreezeThat’s fantastic. Could you talk a little bit about how you upskill employees at Spotify to find new areas of automation?
Sidney Madison Prescott Yes, so there are a few different ways. We typically start with opportunity workshops. And these are, in essence, sessions where we invite our business stakeholders to learn more about the emerging technologies that our team is leveraging. And we also talk through the feasibility of different use cases within their function and we dive more deeply into ways that there can be a potential partnership between the Center of Excellence, which is my team, and the various business stakeholders, specific to the pain points that their teams are feeling. Whether it is lack of data quality, whether it is a need to really accelerate the ability to uptake more processes, more transactions as a result of growth across Spotify, or whether it’s it’s just wanting to again really lift away work that is manual and tedious so that the stakeholders can spend more time on value added task across the spectrum of their functions.
And so typically we start with that workshop. It’s usually a few hours that we sit with the teams, we educate them about the tools that they will have at their disposal, partnering with the intelligent automation Center of Excellence. And then we dive into, again, wanting to learn more about their world. Their pain points, the business processes, and also business continuity. How do we make sure that we don’t invite additional risk into these processes once we automate them?
So it’s a very full-fledged initiative to look across the spectrum of use cases. And then we dive in deeper in subsequent sessions. More specific I’ll say to different subcategories of business functionality within that given team. And we look to see, are there ways that we can actually encourage the employees to build their own automations? So these are very, I consider them, low-code solutions that employees themselves can build out. And so we typically split the use cases into, is it a process where the business user has intimate knowledge of that process, it is relatively low complexity, and does the business user have the ability to upskill in that process? If yes, that process can potentially become a citizen developer project which is then built out by the business stakeholder themselves after they attend a boot camp that we have internally which basically walks them through all of the different technical knowledge and technical skills they’ll need to be successful as a citizen developer.
Jim Freeze I think that’s fantastic. It sounds like a pretty comprehensive program. And at Spotify you see automation as a two pronged approach, attended and unattended. Can you explain to our listeners what exactly that is?
Sidney Madison Prescott Yes, absolutely so attended automation goes back to the citizen developer side of the house. And these are in essence auto mations that end users themselves can build and maintain. And one distinctive quality of these automations, they also leverage the end user’s own credentials. So you can think of that attended robot as a proxy for the Spotify User so that’s one side of the workflow.
The other side are going to be your unattended automations. And these are automations that are, generally speaking, of a higher complexity. Typically these are batch jobs, these are processes that are running globally, they may be running close to 24 by 7. And they typically require, whether it’s additional data reconciliation, data manipulation, there may be a wide variety of steps and complex logic that’s built into these processes. And these are the processes that the Center of Excellence and my team of engineers take on to gather the business requirements, develop, test and then eventually release into production.
And the distinctive value outside of the complexity and I’ll say the level of transactions also is that these are robots that are leveraging their own credentials. So rather than having the end user as a proxy. These spots operate as if they are users themselves. So those are the two segues and we’ve built up quite a significant workflow on both ends of the spectrum. To date, we’ve trained a little over 100 Spotifyers on how to build their own automations. And overall, the the Center of Excellence has released over 150 robots into production.
Jim Freeze Talk about success and starting to get some broad scale deployment, that’s fantastic. Another concept I want to touch on at Spotify is human augmentation. We have a strong point of view at Interactions, that AI needs humans and humans need AI. Consistent with that notion, you’re a strong advocate for human and machine collaboration, something I think you call human augmentation. What does that mean to you and how do you see that playing out?
Sidney Madison Prescott Absolutely. I am definitely a proponent of human augmentation and it’s a concept that I’ve created whereby we are really talking about rather than seeing humans and machines as separate entities, with separate workflows, separate ways of working, error handling, et cetera that we merge the two together. Meaning the workflows themselves almost become seamless in terms of humans being able to complete a certain portion of the workflow, the robot or the machine completing a certain part of the workflow, and then maybe even having that workflow go back to the human for additional work, and then back to the robot. So creating a seamless trajectory from end to end along that process in relation to the human and the machine working together.
So at the end of the day rather than separating the two out, you almost have the robots, or I’ll say the machines, becoming almost an integral part of the team itself. So then you would amplify this by then having virtual workers. And your virtual workers would actually be the robots that are a part of the team, that are helping to facilitate the workflows across the human members of the team. So then it just becomes almost a seamless dance of humans and machines working together across various processes, and being able to again pass those different pieces of the workflow off in a very almost organic sort of way.
I think we are starting to move towards that at Spotify. So we have a lot of processes that we leverage human in the loop capabilities. So again that passing off of certain checkpoints or certain parts of the workflow between the human and the machine. I think the next iteration of that goes even further, where we start leveraging more cognition in our robots, thereby creating a dynamic where they can become, in essence, virtual members of an existing team.
Jim Freeze You’re kind of touching on my last question which is you know you’ve talked to a lot about automation’s role in your workforce, and the upskilling you’ve done at Spotify. We always like to ask guests on The ConversAItion to project 5 to 10 years into the future. So what areas of opportunity are there for additional intelligent automation at Spotify for the next five, ten years?
Sidney Madison Prescott Great question. One of the biggest pieces of the puzzle is the amount of time that humans spend in meetings, and I’ll say reading, digesting and responding to email correspondence specifically. And so and I think we don’t give enough credence I’ll say across any industry to the amount of time that we spend in meetings and the amount of time we spend communicating ideas, concepts to one another. And so I think this is an area that really could leverage innovation across the next few years to, in essence, decrease the amount of time that we spend as humans in meetings, communicating back and forth and exchanging ideas, and increase the amount of time we actually spend reflecting and thinking about solutions to the problems that we’re looking to address in meetings.
And I think there’s a variety of ways we can do this. There’s everything from natural language processing that we can leverage in terms of understanding like sentiment analysis on email communication from team members, leaders, et cetera. There’s also artificial intelligence that we can layer over that to understand how machines can eventually respond to our peers via email in such a way where you don’t know whether it’s the human or the machine that actually wrote the email. So I think that’s that next level where the machines understand us and our thought processes to the point where they can help us cut down on the amount of communication, because they’re taking on the bulk of, I’ll say, the simpler conversations that we’re having around: do we really need this meeting or can this be something that we can talk about in email? And can the robot actually facilitate that conversation once it learns the way that I think, the way that I typically respond to different types of communication on an email level.
Jim Freeze Well, you certainly paint a very fascinating picture of the future. It’s been so interesting to hear about the role AI, automation and RPA play behind the scenes of a platform that most of our listeners probably use every day, and might even be using right now to listen to this episode. So thank you so much for joining us. It’s been a real pleasure having you on the show Sidney.
Sidney Madison Prescott Absolutely thank you for inviting me, I had a great time.
Jim Freeze That’s all for this episode, and this season, of The ConversAItion. A big thank you to all of our fantastic guests, and to all of our listeners tuning in. We look forward to you joining us for Season 7 later this year.
This episode of The ConversAItion podcast was produced by Interactions, a conversational AI company. I’m Jim Freeze, signing off, and we’ll see you next time.
[UPBEAT MUSIC]
