The recent surge in GenAI buzz has been overwhelming! This has prompted many people to ask: How can I identify GenAI potential and boost my organization’s productivity without incurring exorbitant implementation costs?
Join us for an insightful webinar, where our panelists will explore:
1. Establishing a Digital Core (Data) for GenAI: There is no AI without data. How to set up a strong, organization wide digital foundation?
2. GenAI Potential Detection: How to identify areas within an organization where Generative AI can have the most significant impact?
3. Measuring Impact: crucial metrics, frameworks, and methodologies for measuring the impact of GenAI. Real-world insights into establishing meaningful metrics to gauge the success and effectiveness of GenAI initiatives.
To download the report discussed during the webinar, click here.
Transcript:
Transcript:
Sarah Burnett: Hello! And welcome to our webinar on generative AI. I’m sure you’ve heard a lot about it already. There’s been so much buzz about the technology. But will it actually pay off? What impact does it have on organizations? I’m delighted to say that today we have none other than our Co-founder and Chief Executive Officer, Adam Bujak, with me, and my very good colleague, Henry Ellender, Head of Sales. Hello, Adam and Henry!
Henry Ellender: Hello! Thank you for having us, Sarah.
Adam Bujak: Hello Sarah.
Sarah Burnett: Great to have you here! This is a really hot topic, isn’t it? So we’re going to be delving into it and finding out how to make the most of the technology throughout the day. During the conversation today, please feel free to post questions on the Q&A tab on your Zoom panels. But to start with, I’m going to talk about autonomous enterprises and how they benefit from generative AI. Then I’m going to invite Adam and Henry to discuss enterprise perspectives, with an introduction to KYP.AI if you aren’t familiar with us already. We’ll explore how to make the most of the technology and then touch on the culture of organizations. We’ll examine how the approach to investing in technology matters. We’ll then go on to a bit of future gazing and talk about the potential for a three-day workweek and have a conversation about that before we wrap up with our Q&A session.
Henry Ellender: So for Sarah, who is being modest in her introduction there, for those that don’t already know, she is one of the visionaries. It’s fair to say, she’s foreseen and documented the inevitable rise of autonomous enterprises, even writing a book on the topic. So my question, Sarah, to start this off is: How did we get here today to be talking about AI with so many industry leaders?
Sarah Burnett: Thank you so much. It’s changed since publication. Absolutely. Thank you so much for bringing that up. Really appreciate it. Yes, you’ll see the cover of my book on the next slide. I wrote it, I guess, about 2 years ago, researched it about 2 years ago, and it was published in February 2022. At the time, generative AI wasn’t as big a thing as it is today. There were things like DALL·E and Amper Music, which I mentioned in my book. They were around. But if you looked at what DALL·E was producing as images, it was nothing like what it’s capable of today. So, huge strides have been made in the development, and the way that AI, especially generative AI, can help us today to augment our skills is absolutely phenomenal. Just to recap on what an autonomous enterprise is, as I defined it in the book: it is an enterprise that is highly automated, basically. And it uses all kinds of technologies, including all forms of AI, to automate its processes. It’s very data-driven, and because it’s data-driven, it’s also quite innovative because it finds opportunities for innovation in the data that it has. And it’s very agile. It’s willing to invest in technology of the future within measured, controlled environments to minimize risks. But it’s not afraid of taking risks, either. You have to take risks in order to learn and develop your capabilities. To me, the term encompasses generative organizations as well, because generative AI is another form of AI, and it just extends the potential for automation. Autonomous enterprises would definitely do that; they would use it to automate their creative content production, including images. And you know all the other capabilities that generative AI brings. So, in the book, I have this AI fountain, out of which springs all kinds of capabilities and solutions and devices, from the digital world to the physical world, and within AI, its skills and capabilities, the ability to help with decision-making, innovation, and creativity as well. And I’m really delighted to see the last two, particularly ideation, innovation, creativity. These are all coming to fruition with generative AI. With the advances that have been made in the technology today. Should we just do a recap on what generative AI is by talking about DALL·E, and the fact that initially it wasn’t producing anything this good. But just to show you its capabilities, I just asked it to create a little robot in the colors of our brand, and here it is, just a little robot produced on demand and delivered exactly as I asked for. So generative AI uses its training in existing material to generate new material. And accordingly, it can help with creativity, hyper-personalization, and also improve problem-solving, data augmentation, and also produce synthetic data, which really makes training new AI a lot easier, and also helps with prototyping and simulation. So just a bit of scene setting for you. And we’ll now move to the next section, which is going to be, ‘What does it all mean for enterprises?’ and ‘How can they tap into the benefits of this phenomenal technology?’ So I’m just going to bring Adam into the conversation here because this is very much his domain. Adam, would you like to tell us about the impact on the enterprise and how to make the most of it?
Adam Bujak: Absolutely. I’m so excited. You know, Gartner once outlined our platform’s unique expertise in identifying and capitalizing on genAI opportunities while measuring the associated benefits. And the question here is, for everybody, how to make the most out of it, right? And navigate the ROI-driven adoption, becoming either a focused champion or a calculated spender. Alright, we will try to clarify that in the next section, and we’re looking forward to it. Henry, tell us more about Gartner.
Henry Ellender: Certainly. So, of course, we started out on the generative AI topic at about the same time as everyone else. Probably even Gartner didn’t have that great a head start. But we’re very happy as the months pass that we’ve deepened the way we can support customers with our technology, and Gartner has recognized that as well. So I’ll leave this up on the screen without reading through it, for people to capture. The key difference here at KYP.AI is we had, and we hold, a unique position to be able to understand the breadth of work and the detail of productive work across the enterprise. That means we’re fantastically positioned to understand, not just productivity or trends, gaps that technologies could fill, but really opportunities for technologies, such as generative AI. At the high level, this evening, everyone’s read the quotes at the high level. It looks like this: we turn activity into data, productive activity into data. And from that, we generate insights, generate actionable recommendations, where you, your partners, technology providers, would plug the gap with a solution once the need has been identified and the business case generated. But of course, transformation doesn’t stop there; the results, proving the impact, the return on investment, actually measuring the outcomes of the project, feeding that back in and figuring out the next best opportunity, is what KYP.AI can help you map, describe, and deal with in real-time. So that’s quite high level. That’s the theory, Adam. What does it look like in practice?
Adam Bujak: Well, KYP.AI stands for ‘Know Your Potential’, right? And let me start with one incredibly important statement: There’s no AI without data. Cloud migrations did not fulfill their expectations. You know, we were hoping that everything would be available at our fingertips. What we provide is digital core creation, the data for AI. We deliver incredibly sharp, data-driven opportunity identification, basically answering the question: Who does it apply to, right? It does not apply to everybody, but who does it apply to? And obviously, data-driven impact measurement, which is key in this exercise.
Sarah Burnett: And what about going beyond the buzz, Adam? And generating value out of it? Absolutely. Everybody is speaking about Chat GPT.
Adam Bujak: This is a topic that, you know, I’m absolutely fascinated with—the intensity with which we all embrace it. So, I’m really excited to speak more about that. Now, when you look at one of our supply chain customers, we see that ChatGPT is here, whether you permit it or not. Right? Your employees are using it on all their devices. The customer I’m talking about posted more than 12,000 queries in the last two quarters. Right? They are related to individual challenges and knowledge gaps, like, ‘How do I calculate the first visit in Brazil?’. There are always lots of personal questions. For instance, ‘How do I prepare for an internal job interview?’ So, it’s very exciting what’s happening here.
Sarah Burnett: Adam, if I could just ask you about the value again. The question is, you know, how do we make it worthwhile to embrace the technology? How do people embrace it and make the most of it?
Adam Bujak: Absolutely, ChatGPT users are more productive. There was an assessment that was conducted as a questionnaire by Harvard University and MIT, stating that BCG consultants are up to 40% more productive. We measure the real data, right? And the fact is, what is illustrated here, for instance, is individual team members when compared to their peers in the team who use ChatGPT intensely are more productive, right? This is a data-gathering process example, right from one of our customers. And we see productivity uplifts between 10% and 15%. So it’s not the BCG result, the magical 40%, but things are moving, right? The question is where to apply it.
Sarah Burnett: And it’s not just Chatbot AI, is it? There are many other technologies. What other technologies are there that people should be thinking about?
Henry Ellender: It’s, of course, not just chat. Google Duet, Amazon Q, ect. Many organizations are dodging specific tools, and they’re trying to understand patterns of work that they could then figure out the tooling for later on. So, where is my free-text entry happening across the organization? Where is creative working time happening? Where do people spend the most time searching for knowledge as well? But the trending topic, it’s fair to say, and as you’ve already guessed from the slide, is Microsoft Copilot. It’s so easy for big organizations that already have users of the Microsoft suite, and who clearly find the security discussions around another Microsoft product to be much easier than some of the alternatives that we’ve listed.
So perhaps then, Adam, it’s a good time for you to help us understand if Copilot is going to improve. Every Microsoft users productivity, and if productivity alone is what we should be focusing on.
Adam Bujak: Thank you so much. I love to use the ‘brain map’ analogy. The left side of our brain is the logical part, focused on logical reasoning, and this relates to workforce productivity improvement. There’s also the right side – the creative one, right? How do I establish new ways of working? I may not attend a meeting where 20 other colleagues were present, but I can very quickly see the summary, the actions, and also navigate all the details. Data presentations, excel files that we’re relating to that. So, I am very quickly able to find much faster ways of working, right? And how do I make sure we are becoming more impactful? However, it’s not for free, right? Therefore, when paying lots of money for Copilot licenses, one needs to look at ROI-focused Copilot applications. If we go to the next page, you’ll see the most important aspect from my perspective is creating tangible impact. On the one hand, we have measured, in this practical case of a pharma customer, up to 19% of productivity improvement. And on the other hand, financial returns from Copilot, in this case, up to half a million. I’m going to explain in more detail in a minute. But the key thing is, who is the person in your organization that will truly benefit from it, right? Which processes do I deploy, and to which locations? This is the riddle that has to be resolved, and KYP.AI helps to address it, moving to the first dimension. You know, Sarah Burnett, when she’s asking me, ‘Adam, what are the key factors here?’, I always say, ‘Let’s start with the basics, right? Building a digital core.’ We measure, for instance, the total time spent on productive applications, right? And if, like in this particular pharma case, Microsoft Office applications sum up to 90%, you have to slice the benefits by half, right, because Copilot is addressing only 50% of those particular employees that are executing their very specific processes. Alright, so that’s the first conclusion. And if you want to look into more details, I would like to excited you with much more profound insights. Because, you know, when we meet the CFOs of our customers, they’re not speculating about the huge potential; they want to see numbers, right? It’s about the employee cost, it’s about targeted adoption, and obviously replication of best practices at scale. In this overview, you see two cases: team one generates almost a 20% productivity gain, whereas team two only 4.8%, right? For obvious reasons, we skip those that deploy Copilot but do not generate any benefits, as there’s no point in discussing that. The most expensive teams from the US have the highest chance to generate a break-even, and the focus must remain on adoption percentage and replication of best practices. We help to build a digital core and provide insights showing where you can benefit the most from Copilot and, in this way, accelerate the break-even point. So, it’s absolutely critical to look into these details.
Sarah Burnett: And, Adam, I guess the way that investments are made and the culture of the organization matter beyond these data points, because, you know, it will make a difference in what they achieve having spent the money.
Adam Bujak: Absolutely. And you know, with KYP.AI, we help you to maximize the ROI primarily on two fronts: horizontally optimizing your deployment cost. So maybe at the beginning you need only 100 or 1,000 licenses instead of 10,000, right? And vertically by measuring the productivity impact. Innovators dominate headlines; scalars dominate markets, right? And our focus champions, in our Investment Impact Matrix, need data. Yes, we call it ‘digital core’ to see, ‘Where do I go first to deploy, generating high impact?’ and then replicate and expand with the same methodology across the entire organization. Right? So I have a strong focus on AI-driven results and at the same time meet all the expectations with the excitement. Right? That your CFO will not forgive you if you just remain a ‘cloud expander’, because sooner or later, they’re going to ask you, ‘Okay, I gave you 7 million for your licenses. Have you generated 20? I don’t see them, right?’ And that’s why we say, ‘What gets measured gets done.’
Sarah Burnett: Absolutely. Yeah. Well, thank you very much. That brings us to the last section of our presentation today, which is all about the future, and whether the 3-day workweek is going to become a reality. Bill Gates made the headlines with that when he said it would be okay for people to work fewer hours. If it means working 3 days a week, then so be it. And I guess if people can make a good living while working 3 days a week, that’s the ideal scenario. I can tell you that I work part-time, and I very much enjoy having time to pursue my other interests outside of work. It’s perfectly possible, isn’t it, Adam and Henry? Who would like to pick this up first?
Henry Ellender: Sure, I’m happy to jump in, I think, from my point of view. Studies show over and over that this is seen as a positive step. However, a lot of the organizations that test out shortened work weeks, whether it’s four days or three days, have question marks about how to extract productivity on those working days. They particularly feel the absence of employees on the non-working days when knowledge is no longer around and can’t easily be interrogated. This step to make it happen needs to be about not just putting in extra hours on those working days but really about augmenting the workforce and access to knowledge, Adam.
Adam Bujak: Yes, well, interestingly enough, we see a lot of high-performing individuals who are able to get their job done in three days, right? That exists all over the planet. And therefore, for some people, in my opinion, it’s already a reality. The real question is how to execute it at scale, merging the power of human and machine, right? Because at the moment, it may be a 99% human, 1% machine ratio. Here, the ratio is 60/40, right? And the question is, what are the other ways? And I hope you’ve learned a little bit about how we help organizations do it, how to increase their ratio, allowing humans to focus on more value-adding activities, and at some point perhaps get paid for five days while working only three days a week.
Sarah Burnett: It would be ideal, wouldn’t it? Yes. Just a gentle reminder that you still have time to post your questions. On the next slide, you’ll see that we have published a paper on this topic, and you’re all welcome to download it from the address shown on the slide and the QR code as well. So please don’t hesitate to visit our website and download the paper. I’m just going to check and see what questions we have. But before I do that, I have one question related to the slide you showed earlier, Adam. It’s about the variations in the levels of adoption and what difference that would make if we do invest, say, in GenAI and have lots of people with the technology available to them, but perhaps not using it. What would be your recommendation to enterprises on how to increase the adoption rate?
Adam Bujak: It’s a very straightforward story for me, you know. Nowadays, when we are stuck in a crisis, I congratulate everybody who is enjoying their economic situation at the moment. It’s still very much focused on the bottom line, right? And how do I then stimulate my top line growth? So, therefore, the cautious approach here, where I basically narrow down the areas that fly. But in a data-driven manner, right? Identify the opportunities. You know, we’ve been identifying opportunities and execution patterns in the last years on five continents. And for me, the most important element here is when I know what I have, right? And the human brain is not able to process the complexity, even if you have ten employees; understanding what they are doing is pretty tough. That’s what we see with team leaders who look at the insights we are generating. Sometimes they are quite surprised by the great performance of introverts that never speak about that, right? They love KYP.Ai but in a natural way. It’s data-driven identification of opportunities and then validating. Have I achieved something, right? So I would not start in a big wave of thousands. I would first isolate a couple of teams, maybe across different locations, and then execute it in a very, very focused manner, so that I can replicate it. It requires moving across different cells of the organizations. Right? That’s why KYP.Ai is probably one of the few companies that can scale that right. We do not focus on one part of the organization. We go between the silos; we connect the transactions flowing in between interactions. And that’s something incredibly unique and allows us to add value. Here, I see there are a couple of questions. Alright, so maybe I will just go through them quickly with Sarah and Henry. The first question is, how can organizations identify the most promising areas for applying generative AI to achieve significant impact? We provide a very light assessment, right? We are simple to execute for IT teams, right? There are no complaints, right? Where it gets executed very quickly, and then, you have to search for patterns, right? You’re defining what you’re looking for. In the Copilot case, it’s pretty straightforward because we are focusing on MS Office applications, but we also go beyond, right? So for all of you that are on Google Apps, it’s not a barrier at all for us. It’s also a pattern, right? And then you have to replicate and apply it across. But, as I said, data-driven decision-making helps a lot. The second question to me, so I will take it. Question to Dr. Bujak: Do you see any parallels between the RPA hype a few years back and the current excitement, and any lessons to be learned? Great question. Thank you so much. We’ve all been part of the RPA hype. I have to tell you that, yes, I see there is excitement, but it goes beyond what was there with RPA. RPA was gluing together the complex constellations. Here it applies to every human on the planet. Right, my daughter didn’t necessarily have a clue what RPA is. Now, she’s struggling not to use GPT to accelerate her homework. So it’s fascinating how fast we all embraced it and how much benefit we can generate in the future from it.
Sarah Burnett: Yes, there are more questions. I don’t know; who would like to answer this one? How should we measure the impact of GenAI on our organization? Is increased productivity the ultimate metric?
Adam Bujak: Let me take this one. I love answering questions. So, first of all, productivity metrics have been defined for hundreds of years, right? For example, if you’re processing invoices, you know there’s a certain throughput you have to achieve during the day. For me, I believe that the key and most important metric in the future will be the balance between human and machine. What’s the percentage of work being executed by each? More importantly, I just read a very interesting Harvard Business Review publication on how to assess jobs in the context of genAI. Actually, the approach here is very simple. You dissect the job into different task buckets, and check the task suitability for genAI. If you want to do it in a data-driven manner, it’s accelerated. You can very easily scale up. I’m not saying no to smaller organizations; we do also conduct smaller assessments. But now, this dissection can happen without interacting with a human being, without disrupting the work that teams, organizations, and departments are executing. And that’s splendid, right? I laugh because I love data, and I highly recommend looking at it from a very objective, data-driven perspective.
Just wondering, Henry, would you like to add anything?
Henry Ellender: I was going to add that there’s a related question in the webinar chat, rather than in the Q&A, about human dependence on genAI. Is it going to become an addiction? Are we going to be over-dependent on that kind of technology? And again, I think there are lots of parallels. Let’s reach back into history for productivity. But I think of things like handwriting versus the ability to type things out quickly now in electronic formats. Of course, it will change the skillsets of the adopters. But those that are working in those spaces get hugely accelerated. And those that want to be creative in traditional, maybe old-fashioned ways, five years’ time looking back on this webinar, it’s still going to have a niche skill set, probably with greater value if fewer people are doing it. So, it’s another element; it’s another tool.
Sarah Burnett: Thank you. One more question, which is, for obvious reasons, MS co-pilot is nicely accessible. I’m searching for the cases that I can build an internal LLM. for. Can KYP.AI find us those?
Adam Bujak: Yes, absolutely. Now, you know, interaction is something that we’ve been executing on very intensely. We’ve become leaders in the interaction space, right? So, this is definitely doable because what we do is we focus on identifying patterns, right? And whether it’s GenAI, broader AI, or simple intelligent content recognition, RPA, it doesn’t matter. Internal LLMs have pros and cons, especially from the information security perspective. It’s much easier to execute. We have found ours, right? So nowadays, you can very easily navigate KYP.AI data with dedicated questions that are answered in chat mode. So, very clearly, yes, and we’re inviting you to connect with us.
Henry Ellender: And it’s a great question, from the point of view of some of the feedback that we’ve received from customers today. So, of course, we talk about accelerating creativity in PowerPoint, in writing, or in trying to understand formulas, but the search, particularly in project work, for knowledge that can be supported by these kinds of models, the way you can make information accessible, really seems to be the emerging win, definitely the biggest trend inside Copilot from the applications that I’ve seen. Certainly.
Sarah Burnett: Well, that brings us to the end. Thank you so much, Adam and Henry. Very insightful. Really appreciate your sharing your views on this, and thank you everyone for joining us. And please don’t forget we’ve got that fact-filled, fantastic paper that you can download from our website. So thanks again and join us again in the New Year; we’ll be doing more webinars then. Hope to see you soon. Bye-bye.
Adam Bujak: Thank you so much. Bye, bye.
KYP.ai Communication 
Sarah Burnett 