Transforming the Enterprise for Optimal Efficiency
You think about the past two years one of the most unifying things supply chain difficulties. Every single one of us has been affected by this. So how are you using automation to kind of take on these issues. Yeah I'm a big time. Thank you for for that question and glad to be here. Supply chain in general is very you know in the left right and center. If you look at last year but always being and especially if you look at the pharma companies supply chain is very critical. Right. Because we have not only that we have to manufacture all the medicines but didn't need to be delivered on time to all the patients worldwide. So focus is always being the cost of goods to make sure that we
automate enough so that we leave or keep the costs down for a lot of medicines but then also making sure that we are driving it towards the customers and our patients faster. So a lot of work has been done. As you know this law work is happening in dumps of iodine aspect off there to making sure that we automating every every sensor that are there and on manufactured units and making sure that we are leveraging our data platform that we are building to build that intelligence on it. So we are proactively monitoring all out of all our devices. We are predictably solving the problem before it happens. A lot of work is gonna be done there. But last two years actually because a pandemic has been excellent big time.
Awesome. Thank you so much. And Amy I think it's interesting. A lot of us here are B2B folks. But you know you're working to make automation benefit your customers at DAX. So tell us a bit about that. Absolutely. So I take sporting goods during the pandemic we had to shut down the stores. And it was really a moment where we were grateful for having invested heavily in
technology and the infrastructure because we were able to turn our stores into 800 many distribution centers basically which was again a way to make sure that the product was a lot closer to the customer or athlete as we call all of our customers. And I think the key thing there is to look in advance. Right. It's not a matter of will the rainy day come. The rainy day unfortunately will come in this case because the customers can no longer go into the store. We had to use the the stores as those distribution centers. So leveraging the automation that was already in place ledger leveraging the backend systems and
making sure that the customer or the athlete was still getting the service they required because a lot of us got outside a lot of us did more active things in that period. And now Jay I know that you have said and written about how there were a couple years where automation kind of slowed down. Maybe it plateaued. Not anymore. Right. Things are possible like. Right. Yeah. So hi everyone. Welcome to my hometown Raleigh North Carolina. Great to be able to walk here and see all these folks here coming out. So thanks. So yeah automation has been around as long as like us humans have used tools to make our jobs easier.
And you know right back if you look in Wikipedia at the dawn of the industrial revolution patents on very clever automation techniques like in steam engines making sure they don't blow up or boil over with kind of intelligent gauges and instead. So we've been automating for a very long time. Right up until recently. But I would say over the last decade but certainly over the last five or so years you know a wave of data in this digital world you can't do anything without being digital which is wonderful. Ordering a pizza to telemedicine and everything in between. So how do we deal as humans cannot cope
with the volume of data. But we should. And this is where automation. So automation we start getting stuck because of that. And there were certain things in the enterprise that were just very difficult interacting with with other humans getting broader views across the disparate business and I.T. data pool. So along comes artificial intelligence boom. Now the that little road bump is where we're over it and all A.I. is opening up. A high powered automation is opening up a whole set of new windows of opportunity in the enterprise for automation benefit US consumers. I guess I'm an athlete I would say. Now if by your definition I'm a customer of DAX so I'm an athlete. So it's all we're all to benefit us. Yeah just go to DAX and you're an
athlete. So I love that. And Troy ISE signing goes obviously health care quite regulated. Industry research quite regulated. How can automation help all of these compliance issues all of the processes of that. Yes sure. So seniors health where we're we're doing a lot of work with automation. And with the advent of machine learning and artificial intelligence it's really helped us move the needle. And the pandemic actually has forced
you know for all the bad things about the pandemic. It's forced a shift in the acceptance of technology with. Science base so we're able to use that now when we have had to do so over the past two years using data to make decisions so using that and showing the benefit and the quality that comes out of that you can you get less mistakes through automation or regulatory regulators are actually seeing the benefit of doing so. And it's actually driving a lot of efficiency going forward. And helping our organization get rid of a lot of administrative tasks and letter are really highly skilled resources to much higher value tasks. So it's been very very beneficial. And has there been maybe an advance in recent years that allows you to kind of accelerate the automation of these more administrative compliance tasks. Yeah. So if I take an example some of the things we've been doing at our company if you look at are some of our highly skilled resources like Clinical Research Associates. A lot of the tasks they do are very administrative. So they'll
go to a site they'll have to gather documents and those documents have to be put into what's called a trial master file. So typically they'd be on a site have to scan those documents into a scanner. They'll get them received into their email and then they have to go back and process them again open up each one and actually submit it into another platform. So through
automation we can actually do some sort of OCR actually digitize that document and get the data out of it for further use but also apply the appropriate metadata to those documents and automatically file it in there. So those same highly skilled resources can do the things that are more important like patient care and working with the site stuff. That's awesome. I think we all remember the pre OCR era and it was a dark time. So another question of automation is always what processes are worth automating. How do you really make sure that you are actually making these systems work well and you're not just
making inefficient things move faster. And again I know you have done work on kind of making sure that different automated processes work together. Would you tell us all about that. Yes I think just for us I think what you said is very very critical that doing the wrong thing faster. So it's very important that everybody invest enough time to understand that what they are actually targeting automating and actually is the best way to do that is a lot closer to the business. I think in a tribal saying actually I can relate to it exactly in the attendees or finance or H.R. you can pretty much go across the board business to corporate application and you can find these repetitive tasks that can be automated. Make it fast and make sure that they are being all you know working through the business that they're the right thing to do before you put any automation behind them. Now one of my you know mission of life is that this is a close and dear to my heart.
The whole intel and automation what I'm calling it almost like a intelligent and smart automation for me is if you look at every organization you will have cell service automated platform something that actually giving people to use. Outside of this you probably have something that's working as an automation from from a box perspective something there. And that's something like a process mining aspect of it. And on and on and on. And what I'm driving in GSK and my definition of intel and automation is that how do I really create this connected equal system of automation. How can I really make one plus one is three to get the most value and benefit out of the automation is how can I really take a cell service environment connect that with my process mining do and the outcome. What can be then can create a robot that can actually automate deliver that pieces rather than individually done in isolation aspect of it. So building and I think industry is moving too. I mean if you if
you talk to some of these large providers without naming them you will see that they are thinking that way the same way that how do we really create disconnected eco system of automation to build us magnets. And then we have the enterprise data platform which is kind of state of the art you know a MASH based harmonized it a platform where we can then put all this data out and then mine it and apply it in machine learning and an A.I. aspect of it to get even more value out of it. So that's something that I'm driving pretty significantly in. And GSK and I know it's awesome. Thank you. And Amy how do you think about
which processes are worth automating when it comes to the customer athlete experience. That's a great question. And just to sort of play off what you're saying the last thing we want to do is do bad things more quickly. Right. And I think back to my first exposure to a I where somebody introduced me to the concept of Chihuahuas versus blueberry muffins. I don't know if you all have have done that will excuse you to Google that really quickly because it's worth a look. And it basically talks
about the complexity of A.I. and how hard it is to do some things that humans can do more naturally. We can tell the difference between a Chihuahua and a blueberry muffin. However a ISE script it's incredibly complex to do that. So at Dick's Sporting Goods we'd like to look and see what is it that we can do to create a true omni channel experience as we call it in retail. So we believe in the in-store online and in the app and we're trying to figure out how do we partner your experience with a teammate that can help you get the most out of it. So one example would be in our stores you can go and get these very cool sensors. You can go get hooked up in a system called gears and I'll check your golf swing. I'm left handed. So forgive my
my bad swing there. But they they put you in these sensors. But a teammate guides you through the experience. So it's not just your going and playing a video game where you swing and then it tells you how to get better. You swing a couple times and then data reads whether your hips are out of alignment whether you're pulling back too far etc. And a teammate helps walk you through that. So that's a case where we really try to partner that team and experience to give our athletes the best possible experience
plus an automated experience to be able to do and see these subtle motions of the hip and the swing that an athlete a teammate couldn't exactly see just on their own. So that combination of an automated process plus a teammate we think it's where the secret sources. And that way you don't go down the path of confusing Chihuahuas and blueberry muffins. And do you have a sense when you offer this. You say omni channel experience where it's online this in-store. Does this help customer retention. Well what's the impact. Absolutely because I think we all want the
same experience. We want the joy that we get from. I just want this right now. We want that same experience in store. And we've even done that in the stores with another tool we have called Shoe Runner. So you can go into the store and while you're waiting on a teammate to help you you can take a shoe scan it and it will tell you if the size is there or if it's on a close by store. They'll make a recommendation which is the same experience we've come to expect online. We want our our robot overlords to tell us if there's a better shoe for us if there's a better fit. Where's the closest one. And whenever you partner
that with in-store you know support of customer service you kind of have the best of all worlds. So we want the joy of shopping with somebody telling us. Yeah. That that she is not for you with the same like where where's the inventory. If we can get those two things kind of no matter where you are if you're in your sport playing it you want a better your best. If you're in
the store you want what you want. And if you're online you'd like to know where that that shoe is and when it's going to get there. Absolutely. And Gerri I know you have spoken about there are certain environments that you think with the new era of automation are just right. You know what do you think about. Yeah. So I think when when you think about automation and an A.I. there are amazing tools like process mining tools to really turn the lights on. Like we tend to operate in that in the dark. And if you've been like me in I.T. as long your spidey senses you trust them almost too much. And but getting that data is is
is vital in process. Mining takes the guess work out. And and but but that's just the starting point. The interesting thing is the fertilizer for automation is more diverse data the water to water your automation to have it sprout and grow. Is integration integrating. You said one plus one equals three. So one example how many of you are delighted when you go to the Department of
Motor Vehicle to come on. Well I am not saying here in North Carolina when you engage and the pandemic has only fertilize. This movie is a delightful experience through their conversational way I interface and it's not just a chat box. And think about the fertilizer and the water. Right. So more data is better. I go to register and it says remembering my past history where I didn't I didn't have my car inspected before I went. And I had to kind of like you know go back to go meet my wife. I'm very cheerful bottom dealing with the whole time I am. This time I'm asked right away to make sure I get my car
inspected. But more diverse data. You have a Tesla Model 3. I highly recommend them very economical. But data shows from perhaps Tyre AECOM that you're going to fail your inspection because your rear tires probably need to be swapped out and changed. I've only had the car for. So it fit. It doesn't go back to go. So I go downstairs with it with my phone and flashlight and look at it. Might be your tires. And lo and behold there you know as rounded as they come. So think about that. That getting the more data. But integration because now I'm using third party source sources. And when I
finally get down and do the job I wanted to do register my car it's done. There's no guesswork. There's no go. It's all well done and informed. And it's it's a delightful delightful. The math along with that and when thought was coming. All right. So these are the types of things knowing your processes is a good start but it's not sufficient. You need that fertilizer diverse data. Water it to grow integrating with external systems. Awesome. And thinking about diverse data. Thinking about the data that comes from automation. Right. That we can maybe get new insights that we wouldn't have done by hand. Right. So Troy kind of what have you seen when it comes to the data that results from
automation. What are you excited about. I mean I can go back to the example I gave you before. You know when we digitize those documents that are being collected it's part of an overall process for clinical trials. And that data usually is something we can use to trigger either an upstream or down stream system to do another action.
It also helps us as an example a typical collection document you get would be a CV and it's a required document for a clinical trial for a P.I. or a primary investigator. And if you already have that data there and you go to do another study and that P.I. is going to be working on that study as well. You can flag that that content is already there. So again you just drive more efficiency through something you already did. So it's very very valuable from that perspective. And then if you can take the other data sources that are there that you can mine within your organization from third party data and internal operate operationalize data we can do better prediction as well. So what's your best bet as far as getting sites selected patient recruitment and where should you head from that perspective. And using that data and algorithms you can actually understand
whether or not you're following behind or your way ahead of that curve and make the proper adjustments. So it's very very valuable. Awesome. Thinking about proactive detection. I know you have said that you're able kind of to see these problems before they present themselves when it comes to automation with data. Can you tell us a bit about that. Yes. One of the things that we are applying these this model to Intel and automation including the data platform that we have that has the capability. Right. Is that we need on a large operations in the organization. Right.
And most of it actually outsourced for the I.D. and like everybody else. However if you look at our operations right from modern day to our supply chain and the commercial organization these are very critical functions. So any outages happening either manufacture in sight or happening commercial site or impacting our medical discovery or development is a huge business and patient impact. So we're looking to ways to than mine some of this data that we have and applying the automation to see if we can predict a failure of the system. So an understanding that maybe like you said dire similarly like seeing this this this file system situation or the high utilization of the system or in a specific day and time and something runs together cause these issues that we can actually predict and proactively remove it through automation. So without any human intervention can we apply those pieces and we're seeing the some of the success there but a lot
more work need to be done because it is it is something that actually is not done right. Can create actually much much complex issues in the environment. So we are we are taking it slowly but that's something that actually can really take not only the cost out but also improved the quality and of the lindsay of the system and then actual impact on our business and our patient would be tremendous. So a lot of what is happening there as well. Nice is someone thinking about automation. I think one of the most knee jerk reactions you get is fear right. The robots are gonna take my job right. And especially as
we kind of head into an environment that some fear is recessionary. How do we make sure that automation actually works for workers had to make sure that it helps people. And Amy so how are you kind of hearing this and how are you thinking about trying to make sure that automation works for everybody.
That's a great question. I think we are playing on your let's not do inappropriate processes more quickly and experiences with the DMV or in health. Right. We all want a better experience. We want to be known as a customer athlete patient. I think it's reassuring people that we're going to take the work. You don't want to do the processes that are boring that you have
hoped would not be yours to deal with. And we're going to automate those processes. And one of the ways we do that at Dick's Sporting Goods is everything from hosting CAC fans to find out what problems can we solve to engaging our store teammates to listening to our athletes so you can crowdsource it and then people are involved in the process. And it makes us all part of how we want this experience to be better and look for the processes that are so repetitious that no one truly wants to do it. Nobody wants to wait in line at the DMV. You know no one wants to wait in their doctor's office or for their medicine to
not go to the appropriate place. So when we find that there are places where it's naturally given to human error or. Matter of it's just so repetitious that we would rather something a bot handle it so that we can do something more interesting and a great way to do it is to reach out to your employee base because rather than it make on us vs. them it makes we're all in this together to imperium because our center our center of excellence is about serving the athlete or the customer. ISE. So if I had
to. Absolutely. Well one other things. And I love what you said there. What I'm hearing and that is also trust. And with anything automation or A.I. related that this piece of software is gonna do this mundane job that I've been doing so I can do something of higher value. Who's culpable if it doesn't go right to the piece of software the vendor you know IBM for producing the software. So how do you gain trust in that bot that it's going to do. Now as you get to more regulated industries this this no matter if you can't say no. It was hot. You know we messed up that that prescription or whatever. So trust is key. And then IBM back. You know I've never worked on a mainframe at IBM but I've learned so much from the folks who have. And
they've been doing intelligent workload management on the mainframe. And we learned this little trick from from them and we pushed it into IBM. Automation is the three step trust building process. Step one is to alert and inform which is hey I see something that's worth noting. And so it's an alert with an explanation. So explain ability is the key to trust. Then the next step. Step two is alert. Explain. But then show what you would do to automate it to remediate it. Don't do it but just show what you would do.
Trust is being built and then put down at the bottom of button. Go ahead. Do it for me. Just go ahead. I read it. I get it. Go do it. Now you're training the system as well. The third one is fully autonomous. Autopilot mode which is you regain that trust. Now depending on the importance of that work I hear back lessons
from from the IBM mainframe group that it can take up to 10 years sometimes to build trust for a piece of software that's operating on a mission critical system. Of course you know the DMV bot may not be that well. We'll fix it in real time and use an agile process. But things that get more regulated could take more time to build that trust. So the biggest fear I hear is is is around. Is this thing may or may I take my job. But I'd better do it right. How do I trust it. And I think there is a formula for that. Awesome. When thinking about building trust Troy I'm wondering how you how kind of approach this. I know that you have done a great job in your organization of kind of spreading the good word of how I can help. Yeah. And just tagging on to what each of the three of my colleagues here
have said. It's true. Trust. Trust is key because most people with any change and especially with automation they're assuming they're going to lose their job. And I think I've seen it on one of the other events like this where somebody had said that people don't actually lose their job they go into higher value task is what happens. So the key to it really. And I like what you said. It's about explaining why the decisions made allowing the person
to make the decision as you get to prescriptive kind of outputs. But you need champions in the business the people that you're making the change to. They have to understand why you're doing it what the benefit is to them and what the benefit is to the company overall. And once you have those people engaged and understanding that they're the biggest advocate you can have they will drive that change. And the organization works very very well. Go ahead. No I completely agree that I think the business alignment is so critical that they're looking at the value that coming out of the way. There's a cost time or quality but there's really depends. Those three things. But even for the employees I think is very important and very
purposefully that you do a school transformation implementation the organization. You can't sweep away something that they do and then give it to somebody else to do it. I think every time I think that's one thing as industry leaders there's something missing that and we're doing that in GSK. I think pretty much everything and anything that we're trying to bring in that we purposefully drive the school's transformation like up skill like the point steered to people like this is where you need to forget about what you've been doing this last year. This is what need to be done and kind of up your skill from doing the mundane task to actually working this soft engineering to automate these pieces. So it is very important that the employees understand and everybody line up behind them that they see the value as well along with the business. Otherwise you're going to have a
lot tougher and difficult. Glen. Absolutely. And Amy you know when cash is tight obviously it was not a super easy year for retail. Laughs especially 20 20. How do you kind of advocate maybe against internal skepticism that this really is going to help. It just absolutely playing off what you were saying. I think
it's a matter of change management. I think that you a invest in what you save and these operational costs and innovation and you invest for the future. So Dick's Sporting Goods that example I gave right off the top the fact that when the pandemic came we had to shut the stores down because of those longtime ever time investments we were able to rely on automated processes and not only people. So we were able to again in partnership people plus process led us to great stability and to be able to stay in service of our our athletes. So I think when you when you build that sort of trust and the other investment even when times are tight I would advocate for a change management practice within. I actually as part of my job function run change management within a tech organization and people may think oh soft skills you know amongst the amongst the hard coding skills but that that combination is key. It teaches us how to build trust. It creates those champions. So I think that even though these
things may seem extra or soft or these other words that we use to budget cut and set them aside they can be the most important things because it prepares you for those harder times so that you can deliver value through it. Jerry one on one technology and automation that's exciting you the most. What's emerging that you think man this is awesome what people need to know. Yeah. So one. OK I mean I actually I really enjoy looking across you know I said one plus one equals three. That scenario and and for whatever reason business automation and
I.T. automation there is like a great divide between the two. So we're doing work around this topic whole biz ops or business operations which I think really excites me because it starts to look across both business and I.T.. So a quick example of of what we're seeing. So picture it's it's thunderstorms are predicted in the Atlanta Atlantic Atlanta area. And on the environmental telemetry side a bot is now predicting that this is going to these thunderstorms are going to slow things down in the at the airport. And a business automation bot then who has access to your backend systems sees how many flights are going to be delayed and how many passengers. Right. So that is not passing it down. An automation pipeline enriching it with insights and data to a and and S.R. Reebok that then goes ahead and says you know there's a 200 percent rebooking demand I
predict. So let me start automating some writing of service now tickets to deploy some new instances of rebooking up. So what you have is prediction of whether increased capacity for your rebooking application on demand result happy customer all as happy as they're going to get. Right. It would have been worse on that. But maybe you can now offer them with that advance notice some some free dinner tickets or more et cetera. So that bridging the gap between business I.T. and the machine learning
models that are looking across business telemetry and I.T. telemetry I think is one of the most exciting things to come up because it's it's there. That's the next bump in the road. And I think A I can could could flatten that bump out too. Well that's awesome. Thank you all so much. Go get a hand for our panelists.