#11 The Automation Quotes Episode

Welcome back to another episode of the Process and Automation Podcast with the Automation Guys. For today's episode, we have handpicked a couple of interesting automation quotes to see what we can learn from them and what we can apply in our automation projects today. Hello everyone to our today's episode with the title 5 automation quotes and what we can learn from them. My name is Sascha Kutsura and as always with me here today is Arno van Rooyen. Hello Arno.

Hey Sascha, how you doing?

I'm very good, thank you. So today we have a very, very good topic, something different to what we usually do, not too technical, heavy technical focused automation quotes. Some quotes I see regularly and I thought, yeah, it will be a good, a good chance for us to go through some of these really good quotes and see how applicable they are today.

Yeah, some of that's quite a, I guess, a lighter touch to the heavy-duty automation topics that you always talk about. So for listeners out there, I think this will probably bring out a lighter sort of angle to automation. We can discuss it. I think a lot of these, these quotes are actually quite true. Yeah, absolutely.

And some of them have been communicated or have been put out there a while back, isn't it? So when automation wasn't as strong as it is today. So some really good predictions there from some very, very interesting people and some very well-known people.

Yes, good. So first one I've got is automation may be a good thing, but don't forget that Began with Frankenstein, and this is a quote by anonymous person. So what's your thoughts on that?

Yes, obviously. So sometimes I as well rephrase it, raise automation to magic. So, but so you need to really understand really what automation can do for you, and you really have to look into it, into every little aspect. But if, if you don't know and really understand what you're doing with automation lots of things can go wrong. So something was a bad process, if you start automating that one, definitely it will be an automated bad process. And suddenly if you do address changes or so, an automation suddenly makes 10,000 records and updates wrong in systems, and that could be quite horrific. So I think the Frankenstein comparison here Yeah, definitely, it's true.

Yeah, but I think also it's about experimenting, and, you know, sometimes you just have to throw yourself into the deep end, um, you know, to actually experiment with some things and see, you know, if I put these things together— as long as you fail fast and you don't create lots of problems And I think that's okay. I think a lot of, a lot of research into a lot of areas in the world, you know, starts with experimentation. And sometimes you have to create a Frankenstein. Maybe only you will love it because you, it's, you're the creator. Um, but at least it's something that you could look at and say, well, things can only get better from here on, right? That is the worst thing. That could have ever happened in terms of an outcome. You know, a few small tweaks and this thing looks and behaves a lot better. And from there on, you know, things can only improve. So your Frankenstein can turn into a beautiful butterfly or something, right?

Yeah, we interestingly say, so if something can be automated, we really say it should be automated these days. But yeah, definitely it's key to look into it and And yeah, try, try out things indeed and really see if it works and produces the benefits you need. Okay, the next one I have here: automation is cost cutting by tightening the corners and not cutting them. What, what do you think of that one?

I think that's an interesting one. I think let's you know, look at it and don't eradicate stuff from it, just look at it with the view to keep things inside it, but make it better. And I think what this tells me is the person that wrote this quote is probably a person that's a details person, that you know, the devil is in the detail, right? And, you know, I think if you, if you go into automation initiative and you start cutting corners because, hey, it's automation, it should be easy, right? We will just do this very fast. We can, we can cut all of these things out. And then once the automation is done, it will fail because there are certain things that you need to include in the overall automation to make it meaningful. Yes, there are probably certain things that you can remove, as long as removing it still, you know, results in something that's meaningful and works more efficiently.

Yeah, absolutely. Yeah, it's more really fine doing things a bit more efficient, isn't it? And, and yeah. And not going really the drastic route.

Yeah, yeah. So here's one for you. If you're looking at automation as a way to cut corners or avoid certain processes, your outlook is all wrong.

Yeah, that's an interesting one.

So this is an anonymous quote from, from some, from somebody. And I think this probably goes hand in hand with the previous one where it says, you know, don't cut the corners, just make them tighter. So I think this is probably to do with inclusion again of everything, but make it as good as it can be.

Yeah, it's all about stripping out everything what's unnecessary, but still keep everything in there what is really needed. Yeah. And still, yeah, still to make sure you're not overloading here your people. I think that's the big topic I've learned in automation, to really look at the people in the process and make it as easy as possible for them. But otherwise, do not really cut corners too crazy, so your process is disrupted too much.

I think also You know, especially now with a lot of the modern technologies, that's emerging automation technologies, AI, chatbot, you know, RPA, machine learning, process mining, you know, a lot of the vendors do promote the products. And of course, you know, these things are progressive. It does help businesses, but We sometimes get people that surprised that you still have to map out processes. You still have to go through that due diligence to, to sit down and understand things, right? You can't just go ahead and say, okay, you know, we've got AI now, we're just going to plug this in and this is just going to revolutionize the process and cut all of these things out, you know. If it was easy, it would be fast and everybody would do it super quick. Luckily, we do have the tech to deploy automation quite quickly, but you still need to look at the end-to-end requirements and process. You can't just start cutting it out. And I think some people, you know, still are surprised to say, oh, so we still have to think about the processes and all steps. We're like, yeah, you still have to do that.

And because they don't, they don't like to do this, I think this is why process mining is very successful these days, because that, that has sort of the potential for businesses to, to, to uncover the as-is process instead of you have to map it out. But, but still, it doesn't take away that you have to then reflect on the outcome and the output and then do something meaningful with it. But this is probably one reason as well, so automation can do lots of stuff very fast, but you still need to do stuff you don't want to do. So take process mining to find out again how your processes are really running. So yeah, it all comes to that understanding really.

Yeah, so we've got one from Bill Gates.

Yeah, this is one of my favorite quotes really.

So he says that automation applied to inefficient operation will magnify the inefficiencies.

Yeah, that's a bit what I mentioned as well. So what I mentioned was the Frankenstein one. So if you do something in automate with scale wrong with automation, then you will have a very, very bad result potentially. And here's the same, it's what Bill Gates says, and really makes sense to get that process right. And yeah, and do not do unnecessary round trips on the process. I think with automation, yeah, it will be super easy and will not immediately be noticed. Only if you change data wrong with 10,000 records or so, then everyone will notice very quickly.

Yeah, well, I think that the key here is, um, don't automate broken processes.

Yeah.

And it's the whole garbage in, garbage out thing.

Absolutely.

I think you, you would just by automation, uh, create problems faster if you do it wrong.

Yeah.

Like you say, you know, you could automate something, um, using a robot and just create garbage 20 times quicker than before. So I think it is something that you need to look at. The types of processes that you automate, are they efficient? And, you know, spending time up front just to look at those processes, removing some of those inefficiencies, I guess, coupled with the fact that the technology to automate these processes are quite mature, usually would give you a very good result, even if it's your first time that you do it, that you embark on an automation project. So I think if it feels wrong, it probably is wrong, and you probably need to sit down and look at it and think, is this creating more work or is this actually reducing work?

Yeah, I think we had this whole problem in the early phases of getting processes, moving processes to digital versions. So very often we looked at it basically, oh, we took at this manual process, this was a form. So now we put sort of the same thing into an electronic form. It doesn't make the process better. It is now a digital process, but very often the process was just, yeah, it was broken. But that was very often our first attempt back then. And nowadays you will, with the combination of this process mining as mentioned earlier, you will come up with ideas which probably not even need automation or even need process workflow and all that stuff. So sometimes process can be fixed even without technology required. Once you know what is wrong. And, um, yeah, this is, this is a good one.

And, um, so here's another one as well, and I think it, it sort of goes hand in hand with, with the previous one, the automation of garbage. And it says, um, again, this is by Anonymous: automation isn't a silver bullet and it won't fix your broken processes for you.

Absolutely.

And I think If you look at automation as a whole, you can automate broken things and create more faulty results if, if, if, if you do automate that. But, but I think this is where user experience or the term user experience was born, where if you look at a user journey and you've got an extremely noisy and complicated and broken process, and, you know, you involve people that interact with those processes, that's generally a cue for you. If those people are discontent with that experience, you have probably implemented something too complicated or that's broken. I think this is where user experience journey mapping is quite important. So if you have a, let's say, customer care, for example, of course there's a lot of processes in customer care that you can automate. But I think if you look, look from a, from a journey perspective, map all of that out, look at how that journey looks like, how that journey will feel. I think that's generally a sort of a good yardstick for you to be able to determine Is what we're going to do here actually a broken process, or is this something that's going to end up in the end users being delighted?

Yeah, yeah, very good. So the next one, this is an interesting one we picked. As machines become more and more efficient and perfect, so it will become clear that imperfection is the greatness of man. So this is is a quote from Ernst Fischer. So yeah, this is, this is interesting. Yeah, AI really makes everything seem like perfect these days, isn't it, Arno?

Yes, that's right. I think it is a sort of a go-to thing, you know, but I also think it's in the, in the eye of the beholder. And I think, you know, if you look at You know, humans, I think we are full of imperfections. Yeah, I think that as we deploy more AI solutions, you know, that will probably become more evident. And I think that it's probably okay that we have got imperfections, but I think what we should do is just focus our our energy and our time in doing the right things and leaving the automations where we're likely to make mistakes to something that's a lot more capable of making that.

Yeah, high volume, isn't it? High volume work, really manual stuff, 6 hours working on spreadsheets. But of course we will make mistakes. But if that is something we can, with this high volume, give to an AI, yeah, perfect. I absolutely agree.

And also, I think, you know, humans are good at building relationships and doing the things we do. So if we can remove the robotic elements out of humans, the repetitive type activities that these automations, you know, are designed to do, then, you know, perhaps humans can focus a bit more on creativity and creative sides. And use that to our own advantage.

Yeah, yeah, I think that's the stuff, isn't it? So we can solve all these complex things very easily sometimes where the AI would struggle because, or it would require a huge amount of training to cover all these complex elements. Sometimes you look at the little legal phrase, we exactly know what's happening here and the AI will struggle to find the right answer. So yeah, absolutely, let them do what they do best. And end-to-end processes these days are running best where we really select what the human is doing best at a specific step in the process and what the NIA is doing best. Then we have the best results, really. So then there's another one we noted down: humans like dealing with humans and all the non-robotic humanity that comes with them. I think that's clear. So what we said already, isn't it? So we really, really, yeah, we like to do human things and we don't like to call a call center with an AI and, and is asking really stupid mechanical robotic questions, isn't it? We, we can't really deal with that very well and we get frustrated.

So yes, yeah, and I think like, you know, no matter how clever these automations are and appear to be, there's always going to be the need for a, you know, for a human team member or a human intervention to to jump in and be available, right? Again, interesting quote.

Yeah, I think this is the problem, these machines. So when I speak and call them, they can't handle sometimes my German accent, and then, yeah, it's getting— it's really, really getting frustrating getting through. And some of these companies, some companies are really strict You can't get anywhere, you don't even get normal customer service these days, or rarely. For example, DHL, so if you track a parcel, if you don't have the number, that's it. And if they can't understand your number while saying it, you're screwed. Yeah, and I think the AI these days, I think there's a really good aspect of the Google, you probably have seen the videos on Google Assistant, where the Google assistant is actually taking over your little duties, maybe making a hair salon appointment or ordering a pizza or making a restaurant reservation, that kind of stuff. So the Google Assistant now is even going that far that it puts in, when it calls the hair salon, it really has these filling words like, um, uh, so these kind of words in there to appear more human, to appear more imperfect actually. So So this is what the realization is these days. If it's AI, it should behave like a human to be successful.

Yes. Yeah. So what else have you got for us?

You cannot endow even the best machine with initiative. Walter Lippmann.

So I don't know who Walter Lippmann is, but I think what he's going on about is that automation itself cannot think for itself or decide how to complete tasks. It does, I guess, do what you tell it to do. And there's sort of no flexibility. I guess if even if you have complicated AI models, you still have to train it. You know, it's not going to solve these problems on its own. And also it won't recognize, you know, broken processes or anomalies within processes. Yeah, so quite, quite an interesting quote. I think it sort of brings us back to the fact that it is still rules-based, even if you have neural networks, which is, you know, not strictly speaking rules-based, but it still works on probabilities that this is the sort of decision I need to make. And you know, you train those models using data, and if the data is broken, then the decision-making will be broken. Absolutely. And you know, that's quite a tough thing. And I think it still, you know, it still would require a human to look at it and say, did this really do what I asked it to do? And if there's a problem, it can't learn from its mistakes. Ultimately, that needs to go back to a human to, you know, to make a decision and fix the problem.

Yeah, I think the benefit of machines is that it can, based on that feedback we give as a human, how to deal with a specific situation, that flows in this big data database where then automation and with the computer power of the day can come up with an answer very quickly going through this vast amount of data to figure out what was the feedback for this specific issue someone solved and then combining these things. Obviously that appears maybe quite clever, but it is effectively just a series of very fast selects into data, isn't it? And coming up with an answer, but it's all coming from from us. Yeah, maybe it's not really, really intelligent as we are, but it's evolving, isn't it? So now with the cognitive, cognitive RPA and all that stuff, how clever is it really without good data?

Yes, and like I said earlier, I think if historic data is to go anything by You know, you have to have really, really good data, and all of the processes— or sorry, all of the mistakes you've ever made in your life would probably be embedded somewhere in your data. So I think this is sometimes where data can, can be quite dangerous unless your data is, is really good and, and really pure. So I think human in the loop is, is still something that's that needs to be, I guess, maintained, you know, to ensure that we have got that, the eye on the automations, make sure it does what it's supposed to do, just to make sure it doesn't run away and do all sorts of things that you don't understand. And it's very hard then to unpick when it's made those mistakes. Thousands of thousands of times over. I guess it's like a production line where you manufacture cars. If, if there is a problem, you need to stop because otherwise it will just make thousands of, of defective products.

Yeah.

So I think there's always a place for, for humans, um, you know, to, to, to be in the loop, um, to, to to make decisions where mistakes are made. And again, I think I see this also as, in an ideal world, we would have 100% automation. Everything will be fully automated. It's sort of a lights-off approach and we can get on with our lives. We can do things humans like, but I think The reality is that automation is always categorized in groups of automation where you've got your intent is to automate 100%. The reality is you can probably automate only 80%. So that 20% is still going to involve the remaining 20% is still going to involve some human interaction, which is good because those are probably some of the fringe cases, the edge scenarios where it will cost a lot of resource to automate and it might not be very predictable. You know, it might be a very ad hoc process that comes in now and then. Why automate it if a human is dealing with it? Because it's something that a human would be able to deal with very quickly because, you know, you can interpret something and say, oh, you know, I understand where this is something quite complicated. So that's in the 20% box. After 80%, you could probably say, well, if you sort of implement 80% automation and 70% of those can actually achieve 100% end-to-end, you're on a winner, right? So because, you know, if you can automate 70% of things, uh, 10% is, is the exceptions on, on that sort of 70%, which brings you to 80, and the remainder is, is manual processes. You freed up people to only deal with 20% of the really, really difficult things. And I think, you know, that's a very good benchmark to test yourself. Yeah, I've actually succeeded.

Yeah, if every business would achieve that, I think it would be amazing.

It will be amazing. And I think this is, this is part of the thing that we always try to explain to customers or prospective customers. Is that yes, it's great, you can do all of these automations. If you can automate 70%, then, you know, you should consider that as a really, really good success. And then some people will say, well, you know, why not strive for 100? I was like, well, you can, but it would take infinite amount of time and resource to do that. And it's simply just a theoretical achievement that you can that you can get. If you can achieve 70% actual measurable automation, you're absolutely top of your class, right? Because that is a really, really high degree of automation. And that remaining part is probably the processes shouldn't be automated. It's probably something where you need one-to-one human interaction because it's, I don't know, a high-profile case you're dealing with, high-profile account or something like that, where you'll just piss people off if there's a robot, a chatbot trying to engage this contact and talk to them.

They'll be like, no. Yeah, great. So yeah, we said we will cover 5 automation quotes, but let's overdeliver here. I have one more on my list here. Which I really like: automation is good so long as you know exactly where to put the machine. So this is a quote by Eliyahu Goldratt. So yeah, so this is a good one because it's all about really identifying the right processes to enable organizations to thrive. And automating the wrong process Yeah, so we covered that already, isn't it? So, but, but it's a good one, which, uh, again comes very similar to, to the Frankenstein one and, uh, to, to the Bill Gates one.

Yeah, I guess it's, it's a bit like building a beautiful car and certain elements of that you, you still want a human to, uh, you know, stitch those leather seats. So you can see some of the imperfections because it gives, it just gives that character. But some of the other things you want to automate because it's sort of hidden away and it's something where you could argue that, well, this doesn't really add anything of value if this person creates this part because it's somewhere hidden away in the mechanics of this car, right? So I think it's It's sort of that, that human touch sometimes is needed to do certain things, whereas that machine that does the automation, let it do the heavy lifting, boring stuff that nobody really cares about. It still needs to be there, but it could do that at industrial scale because it's sort of somewhere in the gubbings of this, this, you know, this bigger picture. So that's kind of the sort of the way I see that. Put that automation machine where it should be, those boring bits, you know, but that needs to be precise, very, you know, error-free. And again, back to the car analogy, it's those precise engineering elements that you want to get that automation sort of machine in to create that sort of precision and do it over and over and over again.

Yeah, and I think that also goes down to really finding the right process. And so we talked about this as well in our Center of Excellence episode, to really do lots of these pilots, isn't it, to really figure out which process has then the potential of doing I think really good things within the business. And I think failing that experimentation and really finding the right process, I think, yeah, that really, I think, will immediately sort of, yeah, destroy sort of the users' expectations, customer expectations in these technologies. I think it is important to find the right one and as well to kick off and scale your RPA journey. Because very often what we have seen, we pick sort of the low-hanging fruit sometimes, we concentrate on that one, but sometimes this is not really the good one to excite everyone. And it seems like an easy one, but who cares if this spreadsheet is automated? If it doesn't have that impact, no one will then say, okay, now scale this thing and we will fill up our 12 months, 36 months kind of RPA pipeline. So yeah, that's very important to, to know where to put the machine.

Yeah, true. And also I think looking at the types of automations that you, you do, I think you, you need to look at the touch points with, with other humans. And, you know, if I look at, say for instance, a customer onboarding, I want to have that as human-centric as possible. Yes, there could be a lot of automation behind the scenes, but if you deal with a customer, for example, you know, you want them to feel that they are actually speaking to another person, and that person actually provides them with an experience, you know, if it's a customer onboarding experience, that's really good. This customer should know that they feel safe. They know that this process and this person they're dealing with is very good. They know you and they can onboard you very quickly. Yes, there could be automation again in the background that supports that whole journey, but they show that there will still be that element where, you know, I think it would be inappropriate to have like a 100% very cold automated customer onboarding where this person, I'm just speaking to robots now, and yes, I'm onboarded in 5 minutes and it's just the Skynet experience. You know, how will this look like? Am I actually going to deal with humans or is this really just like a sort of a thing in the ether that I'm going to deal with in the future? Whereas other elements where, you know, human touch is not necessary.

Yeah.

You know, and I think that that comes back to the point, you know, put automation machine where it should be and use that savings you make by automating all of these boring things in actually giving people time to interact with other people and build those human-to-human touchpoints in your process because you free time up from these people. And now they could come to work excited and want to deal with customers and have a smile on their face because they know they don't have to churn through 5,000 transactions going into Oracle and logging transactions in a robotic fashion. Boring day at the office, right?

Yeah, absolutely. Yeah, nice. So this was a different kind of episode. And yeah, I hope everyone liked this. This type of episode. If you can give us feedback in iTunes on the podcast or send us feedback on our website theautomationguys.net. And tell us if you like that episode and if you would like to get more Automation Quote episodes, that will be really good to know. And yeah, thank you very much for joining us today and And yeah, see you soon and let's automate it. Unfortunately, that's it again with this episode of the Process in Automation Podcast. If you like this episode, please give us a 5-star rating and don't forget to subscribe to this podcast so you don't miss any upcoming episode. We hope you will tune in next time and until then, let's automate it.