welcome to the new automation mindset where AI Automation and integration come together successful automation is so much more than technology it's a mindset on this podcast we're here to learn about this mindset from Innovative leaders who actually practice it every single day from Fortune 500 companies to the boldest startups these leaders have reduced cost crafted experience and fueled growth with automation they have transformed their companies and their careers I'm your host Marcus zern and as Chief strategy officer and part of the founding executive team at Rado it is my mission to find these top innovators in AI Automation and integration and share their Journeys with all of you you may notice that this show matches the title of the Wall Street Journal and USA Today best-selling book the new automation mindset by a wado CEO VJ you'll hear references to the key ideas of this book the growth process and scale mindsets throughout the show if you'd like to explore them further be sure to check out the book in hard copy or on [Music] Kindle Ted welcome welcome to the show the new automation mindset so it's a it's a a great pleasure to have you on uh so for the audience so Ted Sheldon he is Bane consulting's Global head of business process redesign uh so really knows Automation and that kind of stuff well the first time we talked you know I started to explain Rado and and Ted was like no you can stop Marcus is like I I know I know what you guys do because Ted for a little while he was the CEO of a company called catalytic that we competed back uh back then with so he knows that stuff and then maybe up front and this kind of brings us to uh you're going to tell us what your uh the picture is uh uh behind you there um so Ted is a real AI nerd and he was way way way before anybody else was so last summer the story really goes last summer I think we were conference together over in Miami and and you were like ai ai ai and and and every everybody was kind of thinking like what does he want with this AI stuff uh and now like a year later I think everybody goes like uh oh my God he was like a year ahead of us so so that that's Ted um again I every time I talk to Ted I I learn something new it's uh and then I would also recommend uh on LinkedIn Ted's writing about uh about Ai and and questions around AI it's uh it's worth every single read uh for sure so Ted welcome welcome to the show and maybe maybe to start start off with I mean you know AI well you know automation well how do you think about these two because you know when you think about it in a in a bigger context it seems to me like they kind of fall in somewhat of the same bucket I mean you know if if if you're customer and you you go about AI you go about automation at the end you want to make your company more efficient you more productive it's kind of like the same goal do you think of them as being complimentary or do you think of them you know maybe is AI superseding automation like what's what's what's your take well first of all Marcus thank you for the very kind introduction um and also for having me on the show today let me let me start with uh a little bit of background on how I got involved in automation AI because I think it helps answer the question of how I think that they're connected you know about 10 years ago uh I was in a Consulting role where I had the opportunity to start trying to implement obviously much earlier AI models than what we're playing with today with uh with generative AI um but computer vision natural language processing natural language generation and what I found was that while um it was a fantastic set of tools for performing particular tasks that there were two challenges of using them one was how do you get the appropriate data into the model that you're working with and then once you've generated Insight how do you get the Insight inserted into a business process or workflow uh and so I actually came at automation through the lens of AI um I found that automation was the way to start organizing and inserting data into models and then once the model ran and gave whatever the Insight whatever the purpose was for the model then how automation again could then take that insight and inserted into the the business process or workflow so I have always seen them as really being entirely interl and you know this latest version generative AI which you know we've all started to get our heads wrapped around uh well most of us have started as of November 30th of 2022 but but all of us have started over the last couple years um it's really just the next flavor the next expansion of the set of use cases that we can use these um uh machine intelligence systems on and they still have the same problem they still need to accept information from somewhere and put information back somewhere and we need automation AI together um I think there a longer answer to the question about does automation replace or rather does AI replace automation um which maybe we can get to over the course of the show because I do think there's some really interesting ways to start thinking about what AI can do for us in the future that starts to really change how we think about what automation even is let's so the one thing I I I mentioned before and we didn't get to the picture uh behind you tell us because I think that goes back to the story that you were ahead of of people on the AI side I think you were playing with Deli right tell us about how this uh what this picture is all about um so so this is one of my early uh attempts at producing something with dolli um Dolly for those of you who don't know is uh image generation technology from open AI um and Bane had the Good Fortune of getting introduced to open AI very early long long before the introduction of chat GPT and and so we had access actually to these some of these models quite early um and so I I gave it the prompt it was something like um uh ex presidents riding motorcycles in the style of Salvador Dolly because Dolly you know Dolly right um and I was just really smitten with the image that it generated so I had it printed on a canvas now it's my home office nice nice now it's I mean it can definitely do uh cool things I mean I've seen you know my the thing is always in my head is my 15-year-old son when he uh it was late at night he realized he had forgotten his biology homework was an essay and uh so we put it into chat GPT just just a topic and we were both surprised you know people always talk about automation like doing more with less I was like man I mean there isn't anything better than that I mean less is kind of this happened in like 5 seconds and then more we both looked at and we're like you know if you submit that the teacher's never going to believe you did that so we had to we had to dumb it down I mean it was definitely more with less um all right but but but but jokes asight I mean so so here here's a question so we we've been thinking about this a lot obviously we're in the automation business we we Embrace AI in a in a big way at uh at workato and and really the rationale is that what we think is we're almost on a on a new wave on a on a New Journey if you want think think about like epidemic right uh uh co uh digital transformation was was key I I remember one customer a brick-and mortar department store chain uh that hadn't invested in online right they were in a bad situation and then they had to digitally transform like really really quickly and so it kind of like really made the point to me that you know and some of these waves there's there's winners and there's losers and you you kind of have to decide for yourself on what side you want to be my feeling is with AI we're ahead of another such transformation wave because if my 15-year-old son is any good indication like if you use that stuff you you might be on the winner side if you don't you might actually be on the loser side competitively uh out there how how do you guys think about this at uh at Bay and like what do you you what are the conversations you're having with with clients yeah well I um I I've probably had a couple hundred now conversations with with Boards of directors with executive leadership teams uh with the chief office officers of all kinds information officer Operating Officer executive officer Financial Officer um uh and I think the fascinating thing to me having spent my entire career in technology um is that this is really the first time where non-it business executives are expressing genuine intellectual interest in a technology right I think they've always recognized that technology could help them achieve particular outcomes and so they've been interested in the outcomes and they typically said I don't really care how the technology Works let's talk about the outcome uh and now now they're saying you know explain to me how this works um and I I think you know I uh the credit for that really goes to how accessible uh open AI made the technology back in November 30th of 2022 um you know putting something that looks like a Google search field on a page and you know type in a and then having it have a conversational response that could remember the threat of the conversation something we really never seen from machine intelligence sort incredibly easy to use and intelligible uh uh interaction that you can have with a machine um and so what what I think that's done for a lot of leaders and organizations is caused them to start really thinking about what happens when I have human level intelligence albeit with some challenges which we can come back to but human level intelligence um four pennies on demand accessible 24 hours 7 days a week 365 days a year um and how does that change all the things that we do and then the the next leap from that is how does it change what our competitors can do and that's when the real fire gets started right because you start saying gee if I can do this you know the folks that I meet in the marketplace every day who are going for those same customers can also do this and so then The Next Step Beyond that is who are the people that I'm not even thinking about as competitors who can do something now um and so I think what that has started is a um a level of urgency with executive teams that they need to understand uh and they need to then uh determine the right path for deployment uh for this technology and and around maybe three key uh areas first is what I call better faster cheaper right there's a whole set of things that we do in businesses that we do in the way that we do them because of an accumulation of behaviors over time and it's a lot of human you know human work that moves used to move paper from one place to another and now it moves digits from one place to another um but all of it can be done more efficiently with machine intelligence and automation the second category is how do I actually add capabilities to my existing products and services to make them better and to deliver really you know an improved experience to my uh to my customers and then the third is what can I do that's really disruptive or what will be done to me that's really disruptive that I need to respond to um and in best uh cases what we see companies doing is taking on all three of those challenges simultaneously which is really hard especially because it's moving so fast and companies are used especially the size companies that we deal with are used to planning on annual budget cycles and now we have something that is disruptive in months and weeks and so they're having to tackle all three of those initiatives simultaneously outside of the normal budget cycle but they are doing it I mean I think there there is a real recognition that this is one of those moments of change that U that's going to sweep through uh through the market and and and pick winners and losers interesting yeah no I think that's what we see out there uh as well so if you think about this powerful combination of AI and and and automation right and you said like obviously AI needs to be operationalized it needs to be brought into into the business um what kind of use cases are you seeing out there I mean we're obviously still very early we're learning I mean there's you know open AI brings out like new features every every week it seems like but uh so I I I don't think we're you know complete at any any stretch of imagination but what what are you seeing what what are the early applications how are how are people getting wins out there uh among your clients well you know I think one of the interesting things is you know automation for the time that you and I have been doing it has most often been applied to back office processes and in this case with generative AI what we're seeing is that the inflection point is with front office applications okay these cases um so you know number one is you know any sort of customer interaction so it could be customer support it could be a part of the sales process um uh the um ability to either augment humans to be able to improve the way that humans are supporting customers or even in many cases to supplant a human in that interaction um that uh that this technology gives companies the ability to do do so much more at scale than they can do with human beings um you you guys just had a really great speaker uh at your event um talking about the transformation of the interaction that they get from thousands of emails every day you want to talk about G talisis yeah Daryl Daryl so he he's actually another guest on the on the podcast as well I I I really like him and I really you know like how he's uh thinking we had a great conversation yeah that that one was um basically get classifying in a way getting a lot of the challenge was that he was um getting lots and lots of inquiries and he had to triage it and it was really hard if you're thinking about like 30,000 or so that he had to deal with and and you he just didn't have the Personnel to go through every single one of them and in in those 30,000 were like I believe like 800 or 900 uh like real orders that you should obviously look at because that's at your peril to ignore if you to ignore those and then other ones that were less important but they had to be routed to the to the right people and he did so with um open Ai and getting it included and even like prep preparing a quote by doing an lookups in the systems uh uh with the right parts and all of that uh yeah for him it was magic it really uh it really helped his business tremendous L you yeah no I think it's it's a great example and it's one that we see over and over again and um so many uh times what organizations have historically done is thought of that customer support interaction as being a cost center and you know how do I drive out volume how do I reduce it and I think that if you then change the cost economics where those interactions are are very very inexpensive that actually you want to increase those interactions that you want to be able to reach out to a customer and say hey you just re received our product do you have any questions um you know whereas I would have avoided that interaction in the past um so I so I think I think this is a great example first of a cost better faster cheaper um but then actually a change in the service offering um to allow you to provide a better more differentiated offering so do what you do now but do it even at greater scale in a faster and more efficient way that then creates more value for your customer um so so that's one one area I think um a lot of the disruption that we see is happening in Communications um so an example is um in the healthc care space um uh one of the interactions that happens between patients and their providers and then ultimately the payers is um that there are um uh what's called pre-authorizations that need to be requested so you know your physician is recommended Ed an expensive test uh that you should go in for um and uh the healthc care payer reasonably because we're trying to control healthcare costs in the world it wants to know why do you need this test is it really actually medically necessary um well now uh chat gbt and other other generative AI Technologies like it can help the physician write that pre-authorization letter in such a way that it will automatically be approved by the provider or by the payer rather right so the provider writes the letter and the payer has to approve it um and so those uh those those communication chains start getting disrupted by the capability of the machine to be able to create the communication um you know the there there was a cartoon circulating a while ago that um you know one employee would say I have these three bullet points you know send a long explanatory message to my colleague about this and a long message would be generated and then you would see in the next frame the other employee getting the message going summarize this message down to the salian points and would go back to the three bullet points um but uh but I but I do think that there's all sorts of these places where companies should think about you know what happens in our employee to employee interactions or our employee to customer interactions or employee to vendor or vendor to employee right and um and how will generative AI uh change this and are we going to have Bots talking to bots uh and and and you know first of all how do we use that to our advantage but also how do we defend ourselves when when the other the counterparty is actually using it to their advantage yeah it seems to me I mean if if I think about myself as a consumer right so the one thing I really hate well I'm I guess I'm impatient right so one of the things like going on the phone and then waiting in a queue that I mean it really gets me going I I hate it I really do so doing that on a on a in you know think about digital transformation being able to do that on my mobile phone or even in a web browser that that just calms me down it's a good thing um it seems like with geni there is kind of this digital footprint that I'm having let's say I'm having a conversation with someone from the company uh it almost feels like the llm then knows and I don't have to say that same thing again like it's it's basically like instead of like having to now uh uh um you know uh issue a ticket or something like that you know you you can do that automatically because it's been said in some conversation so why would you have to now in a structured form uh submit it again right it seems like that that unstructured data that's out there is now added to all the structured data and it just makes things it it it removes friction in in in in terms of of my experience as a customer is that how people think about it or is it you know there other things that you uh that you really that that goes through people's Mind through Executives Minds when they when they think about AI yeah I I I I think there's a um if I go back to the separation between the three different approaches to applying G of AI um in the first case and the sort of better faster cheaper World um I definitely think that the ability to um have unstructured data replace structured data as at multiple places in that value chain um is a critical part of why you're getting um the improved value out of these Technologies so for example um I could give you a user interface a graphical user interface as an employee um that you would navigate through to check a whole set of boxes and you know so I was just working with a medical device manufacturer for example that has such a custom app that they built for Engineers uh to um file change order requests um so you change something on the manufacturing line with a medical device it's a regulated device you need to go through reasonably a whole set of of checks and assessments to make sure that what you're doing is not compromising the ultimate safety of that product or you know safety of the physician or the patient that is going to receive that product um and so you know it's it's actually quite complicated right you've got many many different checks to go through and drop- down lists did I did I click on the right things and instead put a chat ux interface in front of that right you can incredibly simplify it take all that structured information that's on the back end and turn it into an unstructured conversation with the engineer have you considered the effect on sterilization right and have the engineer actually answer in written language right and and then be able to create the backend process necessary out of that unstructured interaction so so so in that case I think the example you give is good the other two cases I think are little a little bit different right I think um when you start talking about you know adding uh a new capability um to an existing product um uh and so um you know an an example uh might be in um the medical space I'll just stick with medical since we're you since I started there uh in the medical space um where I have um a um a patient that is receiving a particular pharmaceutical uh and they have questions about it right can they um take uh a chat ux interface to an interaction with the backend information rather than having talk having to talk to a physician and can I as a pharmaceutical company make sure that I've guided that interaction in a way uh that is um going to make sure that that patient receives the right information right and so there um I I never had structured information right the the guidance about how to use that pharmaceutical has always been an unstructured document um I'll give you another case uh we actually something we deployed uh for a supermarket chain in um carefor in Europe um uh you know instead of just going in grocery shopping what if you have the opportunity to have a conversation about recipes um and so again unstructured information all this you know how do you make a um a ban a flambe um and what are the ingredients and and this is actually a great example too of then using automation on the back end to take the piece of it that's structured information and drop it into in this case the shopping cart right so now I've got my whole menu I've got this fantastic dessert I'm going to make banana flambe um you know I want to buy all the ingredients U well the ingredients then become the structured information that gets dropped into a shopping cart and ultimately delivered to my door um uh so so I think I think that you know the the the interesting case there is not how do I um get rid of structured information um but actually how do I take unstructured information which is the the parlance of the interaction but then extract something structured from it and then in the third case where we talk about um things that are sort of truly disruptive right where you're saying let me do something entirely new for the industry um and and here I'll stick with medical again um uh what if uh we all had a personalized AI physician so you've got your Apple watch you've got your aura ring whatever it is all the devices in your home that may gather information about your health conditions and all that that that information is structured right but now it's going into a model which is producing the kind of interaction that we've always wanted to have with a doctor you know I'm not feeling well today what should I do oh well you've been running a fever perhaps you have an infection you know you know my my throat is sore you know what what should I do you know look in the mirror and see if you've got white postures you know whatever it is right um and and so that that um kind of interaction again is not predicated on converting structured information down structure information um it's predicated on having a knowledge sense of what medical conditions are and how to interpret that structured information but I don't necessarily want to Output anything other than that conversation no I think that that makes total sense and it just makes me think it's almost like it's almost reimagining the uh experience that we have with with companies also I mean it it's I mean it's effectively like you know you you mentioned ordering the stuff there uh for a recipe I mean that's what we we would talk about a to to a friend about the recipe and then we I guess before we assumed we had to then break it down into the into the pieces and Order um but no with AI That's not necessary anymore I mean it's like someone you got your butler who who who who does that because it knows and the interaction is actually very human and very uh uh very pleasant in a way it's uh yeah it's it's it's kind of cool so the the one thing is a little different and it's probably more for the process nerds among all of us I know you have a process redesign background I did some of this in my past um the one thing and this is probably more on the backend side but it it uh it surprised me um because we've always been almost trained or thinking about defining business processes as procedures procedurally you know the steps one step after the next step and so on and then I saw an implementation with uh with Gen generative AI that actually took a completely different approach it used these new functions in uh in GPT and then basically just gave gave um possible tasks and then the prompt was almost describing the job description of that person what the person would do and La behold it was incredible to watch how uh GPT actually picked the right tasks with the right input parameters in the right order almost as a declarative process I don't know you is this is this like out there is this kind of new stuff this this is happening yeah I mean so this goes back to the comment I made earlier about Automation and Ai and the relationship between the two and I think historically we have thought about automation as being um uh a a a very prescriptive task of understanding what the work is to be done mapping it out determining where in that that process map we could insert an automation designing that automation to perform that particular task and then once we've created that programmatic code it then goes it it's tested goes into production and then needs to be managed and I think the future that we're heading toward which I refer to as a future of dynamic and disposable code is one where we instruct the model to perform a particular task it writes it you know writes code right it's still it's still executing a set of steps right but it writes that code in the context of that moment um and once that code is executed it actually will dispose of the code because if it's free to write the code then if the next time you ask me for that task things have changed it's actually better to just write it again from scratch rather than trying to maintain some old code um and so what we'll I think see increasingly is you know first for simple tasks and then more and more complex tasks over time that we don't need to map out the AIS process and figure out how to automate it that what we'll simply do is describe the outcome that we want and the Machine will figure out how to achieve the outcome um the um one example I saw recently was um you know send a collections email to all of my customers who are over 30 days late in paying me and you know how how does a machine go about doing that well you know I use the automation approach and I have to map out every step all the logic of you know who who what is a customer what you know do they owe me money when did they supposed you know what does 30 days overdue mean right um whereas the machine you know is able to to actually interpret that outcome request and formulate the code on the fly to pull the data feed and analyze who's late and write the emails and so all I get at the end of that process is here are the emails I'm going to send did you want to review them makes sense I mean it's mindboggling I I um I just really think about this I mean you know the book right the the new automation mindset uh and we talk about these uh these three things the growth mindset the process mindset and the scale mindset you know everything that you've been talking about makes me think about this growth mindset because uh almost like knowing about how we've done this in the past is almost like a barrier to reimagine what this could actually look like I mean almost holds us back in in a way for sure I I wonder what you see out there I mean when you work with clients um do you see any bottlenecks any any friction points to the adoption of AI what do you think in in terms of this growth mindset I mean do you feel we have to relearn certain things or do you see clients also being held back by the by the old ways let me go back to that example of the medical device company that I mentioned earlier um so in working with the team on how to improve that process the challenge was that they kept going back to their user interface their graphical user interface and so they would say oh well we can add a checkbox we can can add another drop down we can fix this button the description is bad let's fix the description right they were they were very much anchored in the way of doing it in the past and thus the solution was to fix that old way and it it was very hard to get them to break out of that and say wait wait wait wait let's talk about what we want to achieve what would be the perfect outcome you know and and and you know the the Breakthrough was okay now I imagine this is not practical but what if you had somebody sitting next to the engineer who could have a conversation with the engineer and so instead of ever touching a computer the engineer was just able to describe what the change was they wanted to make and that person was knowledgeable about all the things that needed to be done to review that change and could then ask the right set of questions and review that with the engineers so that together they would write the output that was needed for the assessments and the FDA appr approval like yes of course that that that's exactly what we'd want but you know we used to do that 50 years ago or 30 years ago but you know obviously we can't afford to do that anymore like yes you can right that's what now you can now you can again AI will be sitting next to the engineer right it won't be a person sitting next to the engineer right and there it's like light bulb comes on they're like I'm sorry what like the the machine's going to be the person like yeah no I mean we should flag all the challenges right the the machine is a statistical engine that will predict the right outcomes it'll predict it based upon the inputs that we provide um and so in the case of following FDA regulations you need to make sure that you are providing those FDA regulations as a part of the prompt that you are putting into the machine um so that you say following these guidelines what would be the right thing to do in this case right and so there there there are ways of steering the machine correctly um and then when you get the output you still need to correct it right you need still need to look through it and say oh you missed this point um but actually one of the really interesting things that I also um am encouraging my my clients to do is to say use the machine for iterative passes like you get something back and you say terrific thanks for that now please review that and see if you can find places where you can improve it right um even just saying something like you can do better than that try again right you get a you get a better response it's kind of it's kind of weird that way but it it does actually work um so we're just beginning to scrape the surfaces of how machines and humans work together in this new modality um uh but but I do believe if you if you start with the right steerage and you end with the right correction um that you can actually have human level intelligence paired with every employee to get work done in a better faster uh and in many cases an entirely Innovative and new way that's fascinating you know the one the one other thing we talk about the scale mindset you know one of you know when we started workato it was started as a low code technology and and you know one of the um one of the ideas was that we we felt very strongly that something like integration and automation uh it was a bad idea to just do that with three people in a central team for the whole company it was it was much better idea to get uh you know try to get as many people involved as possible not not everybody but you know a broad set of people get them to be part of the team involve them because because they knew a lot more better about the process than than the central team ever will um I wonder with AI do you feel that also that uh that you know you this should be broad democratized uh engagement or how any any any learnings there so far yeah no it it absolutely I I think earlier I mentioned sort of the three parallel paths that we see the sort of best um companies pursuing and and my first one on the list is uh individual work for productivity and for that yes you want to give this technology to every employee um now I think there's a lot of ways that that's going to happen so um I'm sure your company like every software company out there is rushing to add generative AI capabilities to your products and so you know I mean if I if I look at um you know the L we all use Microsoft Office well those that really really try hard not to use Microsoft Office I guess use Google Docs but Google Docs as well will have these tools built in right we're going to have all these generative AI tools right there in our office productivity Suite um and uh and then all the enterprise software uh every single enterprise software vendor is building it in every tools vendor is building it in right so so we're going to as individual employees get these tools whether we want them or not um but I think the challenge for organizations is to help employees learn how to use them effectively um because it isn't obvious to everyone um I think you know PE people have said to me oh well the reason you are doing this so well is you're a programmer I'm like yeah I mean there is a certain truth to that right being able to think systematically um uh as you know and the book talks about that as well um that that system process mindset the process mindset exactly um uh you know that that actually is super helpful um and I think it is in automation always been one of the gaps in having low code tools adopted um because you can make the technology super easy to use but if people don't think about things in a in a in a process oriented way it can be hard for them to use those tools regardless of how easy it is to drag and drop something um but uh and and and even as easy as it is with prompt based uh tools I think just constructing the prompt is not an obvious skill set um and so I think organizations um you know our recommendation is actually there's five things that that an organization needs to do to broadly deploy this technology to employees number one uh is make sure you have a provided a safe place for employees to use it right so don't have them going out to you know it's like it's like saying hey go use Facebook to get an answer to our corporate question right no don't use publicly available free tools right go ahead and pay right the the cost per employee is substantially less than the value created um if you don't believe that start with a pilot start with a a small number of employees and actually test and measure but there's lots of research work out there that already shows that it's going to increase productivity so you probably shouldn't worry about that part of it but so pay for something give them something that is a commercial use secure tool then number two you still want to have um uh terms of use right so you know write a policy make sure that you think through that policy about what you do and don't want people to do with these Technologies um then you need a third a communication strategy right it's great you you created a place for them to go and you have a policy nobody knows about it right you've got to think through like how do I make sure that every employee knows where to go and what the right things are to do and then fourth which is sort of what started me done path is education because employees want help they want they want to know how to write a good prompt um you know going back to my comment about steerage and correction right they need to understand yes this machine can do something for you but only if you as a human are really engaged in it you know don't push a button and think it gives you the answer and you can just send that answer off without reading it um uh and then in order to really encourage culture change uh you need an incentive program so that's number five um uh what are you going to do to reward employees for the right Behavior recognize employees for the right Behavior you know one of the things that we've seen uh done well is um you create a public uh place where people can post hey here's the amazing thing I figured out how to do with chat GPT and then other employees can learn from that and you recognize that hey this person has created something amazing that a lot of other people have benefited from give them a check right nothing speaks louder then publicly saying we're now handing Marcus $1,000 for this amazing thing that he's created and shared with the rest of the of the the company right you'll create real will culture change that way so so those are the five things we we think company should do I like I like that advice I really do um I am I'm thinking about this so so so kind of my one of my ideas is you know after every episode I wanted to kind of distill one nugget and then kind of put these we're going to do 26 episodes so 26 nuggets and and see if we really learn something across all these all these interviews so what's going through my head at the moment is really the your your pieces of advice the five pieces of advice it almost seems to me like you're setting up some HR policies of how you're going to work with these digital brains and so I'm saying HR because I do think think after talking to you we should we really have to rethink our interaction with machines it's like they're almost like not it's almost maybe the better ways to think about it like their colleagues like their co-workers or something so there's policies of how I work with a A co-worker in terms of prompt engineering I think there is uh just best practic is how I manage this as a you know just like I manage co-workers I have a now a set of these digital brains and I got to manage them and I got to direct them and I have to go through an iteration process with them get the you know get the more tell them oh you I think you could have done better here and and teach them and and they're probably going to improve right and so it's a it's almost like and and then we have those interactions that are we almost have to reimagine like uh from the old ways of oh there's a user interface or this is how I inserted data into a machine no I think think of like you're working with a colleague and so these things are now possible and but I think it's going to take some time to uh for people to uh to digest all of this yeah no I think it'll take time but but I but I actually think um you know this is one of the really interesting things about technology over your and my lifespan um we've made technology in ways that are more and more easily absorbed into our ways of working and our society and our culture right so if you think about um you know personal computers um the first generation of personal computers um they are very complex devices um and uh you know very few people actually bought them and used them I mean out of the billions of people on the planet very few MH and then if you fast forward to smartphones right the smartphone became a device because of the touch interface because of the graphical interface um uh you know the the all of the user uh finesse that went into the design and I mean Apple gets I think an enormous amount of credit they weren't they weren't the first smartphone company of course um you know there were smartphones before the iPhone but I think the iPhone transformed the consumer experience of the iPhone of the of the smartphone uh and Android benefited from that but so by the time time we had that technology it was much easier to adopt um and now I think with these AI Technologies uh it's even easier um and so when we drop them into a workplace yes there are things that we have to adjust about the way we work and think about work but it's going to be substantially easier I don't need to learn a programming language right I don't need to learn a whole series of complex steps to get the machine ready for the use for a particular purpose um right you think about all the things that we've had to do with previous Technologies and the bar is so much lower um and so I think it'll be a much more rapid transformation um just the introduction of um of vision um which now is broadly available uh if you're a um plus paying customer of open AI you have access to this today and you can take a picture and uh so you know I'm I'm I'm in an Airbnb and I've got a coffee machine that I've never seen before and I don't know how to operate this happened to me recently and I take a picture of it and I'm like how do I operate this right and boom step-by-step instructions right uh that's and I needed my coffee and I needed step-by-step instructions because I was operating that machine before I had my coffee um but um but so so that kind of of consumer level technology brought into the work Force um means that we're going to transform work super fast um you know I I I think 20204 is going to be a year when we all realize that you know everything has to change I think today people are still on denial I he lots of people going yeah you know I'll wait till it works I'll wait till my vendor delivers it I'll wait till I'll wait till the cows come home you know no 2024 we're going to be like oh yeah about that I think I actually need to totally change the way I work I think it brings us a full circle with this uh winners and losers I I really do believe I mean you know if there were winners and losers in this digital transformation wave I think maybe we'll see this at a grander scale uh with with AI uh transformation because yeah I mean we we talked about this same thing I was uh talking to a distributor uh today electrical equipment and we talked about you know how there's a lot of errors in uh ordering in the in this business uh because it's actually really kind of complicated you know you need to take your part and get to the serial number there's like big manuals and I think today with vision I mean you take a photo of the part and say like I want this I want like 50 of those and and I think it's going to work I mean it's uh and it's so much more high quality in terms of the if the if of the order entry I mean I I actually also believe I mean you were the first one who back then showed me a code inspector you know I think it has been renamed to Advanced analytics um much better name I think I mean I think well done on that one I I I never understood why they called it code inspector in the first place but but but I think that is uh transformative and and thinking of bringing those kind of capabilities into business processes and so on that's also mindblowing uh but I think look I I think we could probably talk another hour or two here uh but uh Ted I wanted to say thank you uh so much that was another another great hour uh I learned something again and I'm I'm glad that it's not just me but the whole audience uh learned something and uh yeah I want to encourage everyone to get into into the AI stuff I think both of us feel you said it right 20 24 people will probably you wake up and go like oh oops wait a second maybe I should have started thinking about this earlier I think the time the time is now so so thank you thank you Ted and I would say like for the two of us to the next time which will happen soon hopefully and for the audience uh thank you for for listening in and um see you in the next episode thank you thank you Marcus thank you all so much for tuning into today's the new automation mindset where AI Automation and integration come together if you want to learn more about the key topics we covered in the show you can find them in the book the new automation mindset by our workato CEO Vella also leave us a comment and let us know what you thought of today's conversations and don't forget to subscribe so you will never miss an episode I'll see you next [Music] time
2023-12-30