Behind the Ops Episode 27 === Intro: You're listening to Behind the Ops, presented by Tulip. Madi: Natan, I have a question to kind of kick us off here. So, I know you used to work in the Fluid Interfaces group, right? Natan: Yeah, the Fluid Interfaces group at the Media Lab. Madi: Why was it called the Fluid Interfaces group? Natan: It was actually called before that Ambient Intelligence and only then was called Fluid and my advisor, professor Padima, started it and she comes from a very traditional artificial intelligence, computer science background, but when she made the switch to the media lab, she got fascinated with reducing cognitive load on people, effectively. Madi: Okay. Natan: And she was, she was talking a lot about, and her research talks a lot about, intelligence augmentation, so IA, as opposed to AI, and intelligence amplification. So, basically, how to... help people deal with the reality that, you know, information is increasing, uh, around us all the time. And that technology in part is like the cause or media, you know, if you want to even generalize it, but you know, humans didn't change so much, you know, evolution is like really slow. Madi: Things change around us though, fast. Natan: Our brain is the same and our eyes is the same and we need the same amount of hours to sleep to function well and so on. So that's kind of like the general idea. And the fluid was, um, how do we find new techniques, experiences, uses of technology to create interactions, you know, generally in the field of human computer interaction that, um, make our life better, allows us to have a more human experiences and like it's, it's all about the interface. So a fluid interface allows you to do that. Madi: I really relate to this concept as like a person who hates notifications on their phone. So if we take cognitive load and apply it to the individual that is me, like I don't want to see the text message bubble or the reminders or the alerts. There's like so many different notifications. Natan: How many apps actually are allowed to alert on Madi's phone? Madi: Only the priority contacts on my messages app are allowed to send an alert and it's silence. So it's just the text alert. Natan: And no news application? Madi: No news applications. Natan: No Twitter? Madi: No. I if I like did not need a phone to function I would it at home. Natan: What about Twitter and LinkedIn? Madi: No notifications. Natan: Madi, you run marketing. Madi: I will look at them when I'm ready to. But I can't just be like on getting distracted by... Natan: You're just more disciplined than most of us. And you're not shy about it. That's what's going on. Madi: Well, I think it's because it's really hard to, to your point, be a head of marketing solving problems if I'm thinking about... person liked this page, thing happened here. It's like really distracting though, to your point, some alerts are important and would trigger an action from me or my team. And so I think that's the other part. Like I really relate to the individual cognitive load and, and what that would do to me and I block it, but I also understand or would pull over maybe some of the research that your group was doing on organizational.... Natan: Yeah. Madi: Cognitive load and how data or knowledge goes through people based organization. Natan: Or, you know, when it comes to a group experience. Madi: Yeah. Natan: Because a lot of our interaction with technology is, you know, we're not alone anymore. I guess technology, mobile phones and video calls and instant messaging, whatever, we're never alone anymore, you know? Madi: And groups, I mean, if we put it in our business context, we have like Slack, send emails and all other types of Salesforce alerts, um, and things like that. Natan: And a bunch of it is automated. Madi: And we, we set it up that way in a lot of ways. Like we want to get the information because we would do something based on that. Natan: We also want to do less work. Madi: Less work like that. Yeah, for sure. And... Natan: But it's kind of like we're paying for less work with distraction. Madi: Yes. Less efficiency and your ability to process... Natan: That's, that's weird. Madi: The information. Natan: That's, we're automating one thing, but it kind of works because like no one's job needs to do that thing that we automated. Salesforce flow or that data analysis, automation, whatever. But on the other hand, it's just like sends information all the time. Madi: I think the thing that you're getting at the distraction part is like the value loss and just wrote automation. Natan: Yeah. Madi: And what you extend over is there's a lot of value to be gained if like instead of you know me checking a box when I get an alert because I know I need to check the box that when the action happens it pushes and does the next step without me needing to do that lower quality thing so that I don't even necessarily need to have that alert or just need to know that it was done and I can focus my or my team could focus their time doing things that require more complex analysis. Natan: Yeah, and we sit in this very nice office all day long. We have computers and large flat screens and huddle rooms and all these things and all that is like helps us like get distracted together and alone. Like in a pretty, pretty comfortable environment, I'd say, and that's like, I guess, modern office life. But, uh, what happens in where people are on, um, you know, an operational environment, you know, labs and benches and instruments and machines. And how do you think that maps over there? Madi: I want to dive into that because we've been using the word automation notification alerts and I think to kind of call ourselves out from like a, a context perspective, I think you and I are often critical of automation in the environment that you're describing, not that there aren't valuable, repeatable automated pieces, but that for a lot of the work involving people and organizations in those environments, that automation can be less effective. It almost has some, some baggage, uh, but here we are today really championing the idea of some types of automation. So I'd love to maybe push you and be like, why is it good here? Natan: We have to like, because we love being critical, you know, it's like, uh, premature optimization is the root of all evil. Madi: Okay. Natan: There's a bunch of people, but I think Donald Knuth, who's a very famous computer scientist, like popularizes, and it's like really comes from like, don't try and like optimize a system before you understand how it operates. Madi: Okay. Natan: And like, it's really important for programming. And I think like also applies to automation. It's like, don't try and automate what you can't do manually for a little bit of time and really understand it. So I think that's at the core and that's like, you know, slippery slope, especially as like we're tied to like what we see is what we call automations, you know, it's kind of hard to automate what has an impact on the, on, on something happening in the physical world, like, you know, stopping a production line, say. Madi: Yeah. Like a machine or a quality issue. Natan: Which is very different than like, oh, I'm not happy with like 15, 000 messages in Slack. Well, you know, which you can mute. But factory stopping. You know, it's a little bit different. Madi: So, the automations that we're announcing the launch of today, one of the big values is helping reduce that cognitive load of keeping track of these alerts. Natan: Yeah. Madi: Do you think it'd be fair to say that this is, like, human considered or human in the loop form of automation? Natan: Yeah, but I would put in the framework that you kind of suggested before there is definitely like the use of automation to reduce the cognitive load on people who are working in, running as in leading the teams working in, also designing and gardening those environment that we call production lines and lab, all these operational environments where you have a mix of people, process, and equipment and all that kind of stuff. So that's really important because like when you automate the grunt work that they need to go and file three forms or like correlate this data or like make an important decision that they actually can have a formula run and check some data that is streaming already because they instrumented the line. I think that reduces cognitive load and the human in the loop part is all about trust because they are there. The humans are there to see, well, okay, I made this rule, but I know why I made it because I made it manually like three times before, and now I know this is taking, so that's one side of the trust. This is a human trusting the tech. But the other side is like, I'm going to not go blind following the automation. But like, if there's like a, you know, in, in lean, there's like a famous concept of like one, two, three, stop the line, which basically means if like there was a quality issue. It happened once, happened twice, happened three times, like stop, find the root cause and like fix it. Uh, if I trust what captures the quality issues, then I might trust an automatic process that would stop my line. But maybe that's so involved that you actually need the human in the loop to act on it. But if your speeds of detecting that specific situation, like vary, then you know, maybe you're less efficient or maybe like you have to do work to get to that level of, you know, lean one, two, three, stop the line. So just as an example, that's the human in the loop side of things. But, you know, you talked about the cognitive load on the org. So there's a lot that automation can do, uh, that we actually don't want humans to do. Boring and mundane things like the stuff we say, you know, you see a robot in a manufacturing facility. What do you say about this robot? You know, usually those are the ones in the cage that are dangerous and say like, well, you know, it's doing the hard, heavy, risky work that no one wants to do. And that's great. That's why we have a robot. Okay. They just need to figure out how to program it, but leave that aside for a second. And it's right, but there are also digital robots. Madi: Yeah, that do some mundane and tedious and mind numbing projects you don't want to do. Natan: They love transacting record after record to the ERP. They just like enjoy it. Madi: I love that the joy we're applying to the digital robot where they're just like, yes, condition, event, enjoying that transaction. Natan: Yeah. Crunch this data and, uh, humans are not good at that. And, you know, there's like a whole industry called robotic process automation, which are companies look, you know, looking at screens, simulating mouse movements to enter data into old legacy systems. And these are billions of dollars companies that, you know, who used to do that? Humans. Madi: Yeah. Natan: And... I don't think humans are missing that at all. And I think companies are very happy, like, buying that type of software, implementing it, to automate, you know, and reduce the cognitive load on, in this case I think it's a perfect example of the hybrid, right? Because no human wants to do that so that's reducing cognitive loads on humans and certainly it reduces the cognitive load on the org and its architecture of information. So automations, you know, the stuff we're pushing out there now is definitely going to do that for people. Madi: And I think to bring it into some specifics, you know, we're thinking about like, how do you improve efficiency and productivity? And that's not just like, oh, this alerted or triggered an action that needed to happen faster. And that was an efficiency. But to your point about like if people are experiencing cognitive overload, keeping up with all of these alerts, they just have more problem solving capacity to do better, more interesting work. Natan: So, you know, so I have like this, uh, a little blast from the past meets like your day to day consumer experience. You ready? So, since we talked about your alerts, let's talk about your inbox. Madi: Okay. Natan: I'm sorry. Madi: Okay. Natan: I have to like reveal a little bit. Do you have like filters or what do you do with your Gmail? Like do you automate anything there? Madi: I do have filters and I'm like an inbox zero person like you. Natan: Most of the time. Madi: Yeah. Depending on how busy the week is. Natan: Yeah. Yeah. This week has been interesting with the launch and everything, but yeah. And another question, do you have folders? Madi: I do have folders, and they're color coded. Natan: They're color coded. Madi: Which is surprising, I know. Natan: No, I, I color code my, uh, my labels. Okay, so we do the same thing. Madi: Yeah. Natan: That's interesting. We should, uh, look at the color schemes. Okay. No, but, like, look, the blast from the past is really simple. There is this thing called BPM, business process modeling. And that's great. Assuming, you know, all your business process and assuming that you think that it would like change very slowly, you're fine and start defining the business process, find systems that would fit in that business process and, you know, automate away, no problem, but the reality, like we're all parts of companies that, you know, just think about why we discuss Gmail right now, because you're doing business process modeling, you're doing Madi's business process modeling, that is how you talk to your team and it's how you talk to other interfaces and if I take that away from you and you can do yourself your own business process modeling, like what is going to happen. You're not going to be a very efficient professional. Your team is going to behave differently and you don't even think about that. And it's like, where is that for people on shop floors that know their stuff the best. So knowledge workers demanded, perfected, and companies answered and various things. We talked about RPA, we talked about Gmail from a consumer perspective. We can go on and on and on. And in a way, you know, that's why I think like rigid business process modeling failed and will continue to fail. And that's really from a very simple reason that the world is too complex to model. So the thing you can trust is humans and their intuitions and their ability to operate systems to put together something that works for them. Madi: And I think this is a great setup for what this like business process modeling flexibility for operations that is automations looks like from the people that built it. Natan: Yeah. Elevate and automate on your own. Outro: Behind the Ops is brought to you by Tulip. Connect the people, machines, devices, and systems used in your production and logistics processes with our frontline operations platform. Visit tulip.co to learn more. This show is produced by Gaby Elanbeck, and edited by Thom Obarski. If you enjoyed listening, support the show by leaving us a quick rating or review. It really helps. If you have feedback for this or any of our other episodes, you can reach us at behindtheops@tulip.co.