Generative AI and Leadership, with Accenture CTO (CXOTalk #795)

Generative AI and Leadership, with Accenture CTO (CXOTalk #795)

Show Video

Today, we're talking about generative  AI and leadership with Paul Daugherty,   Global Chief Executive for Accenture Technology.  My guest cohost is QuHarrison Terry, Chief   Growth Officer for the Mark Cuban Companies. Thanks for having me, Mike. It's exciting to   be able to talk with Paul on AI today. Paul, why don't we begin by asking you   to tell us about your work as the chief  executive for technology at Accenture?  Accenture is a large organization. We're  about 740,000 people, over $60 billion in   revenue. We help companies do amazing things  with technology. That's what we're all about. 

Do you want to give us, to start, just  kind of a brief overview of generative   AI? I think everybody in the audience knows  what it is. But in the context of business   and in our world, where does it fit today? To talk about generative AI, you have to   talk about AI first. AI has been around for a  long time, and all of us use AI continuously.  The three of us talking here and anybody listening  has used AI dozens if not hundreds of times today.   AI has become a pervasive part of our life  through the advances in machine learning and   deep learning and such that have come before. AI, as I'm sure most of the audience knows,  

it's an old field. The term was invented, I  believe, in 1953 at a conference in Dartmouth   70 years ago. And it's gone through  a lot of iterations over the years.  I like to think about three forms of AI: • Diagnostic AI, which is using AI to   diagnose things. Often, deep learning and the  like to look at, for example, using machine   vision to look for manufacturing defects (a thing  we do commonly), to unlock our phones (as we do   every few minutes of every day), or assistive  driving features in cars. That's diagnostic.  • Then there's predictive AI, such as AI we  do to do retail forecasting for companies,   often machine learning and optimization  models. Those are well-established  

techniques. We have lots of people  doing that work for lots of clients   around the world, and many companies use it. • Generative AI is the new thing on the scene,   and it really is a massive breakthrough,  probably the biggest breakthrough in AI to   date. And what we're really talking about  with generative AI is foundation models,  

which are really powerful models that can  be reused across many different use cases.   That's why they're called foundation models. Large language models are a type of foundation   model that really understand language and  have allowed us to really master language   through artificial intelligence. Then the  transformer technology added on top of that  

allows us to generate things. GPT (generative  pre-trained transformer) are these large models   that then have transformer technologies.  They can create new sources of content.  That's really the breakthrough of generative AI;  models that are very powerful and can be reused   rather than bespoke data science projects combined  with foundation models, which have tremendous   reuse and power, combined with this creative  capability to produce language content, whether   it be graphics, video, et cetera. It really is  transformational in terms of what it allows us  

to do as individuals and what it allows companies  to do. But we're at the very early stages still.  Hey, Paul. One of the things that I want to talk  to you about today is the whole concept of you   thinking about this stuff almost a decade ago. In  your book Human + Machine: Reimaging Work in the   Age of AI – sorry, I don't have it in front of me,  but I did read it a while ago – when I was looking   back at that book, one of the things that you  talked about was how AI would ultimately become   the ultimate innovation machine. It's fascinating that it's 2023,   almost 5 years later since you published that  book. What's your take? It seems like you're  

spot-on, but what things happened in generative AI  that you didn't envision or forecast back in 2018?  I think the premise and all the precepts in  Human + Machine really have stood the test of   time well and the concepts we talked about,  the human plus machine, and the idea that AI   gives humans superpowers to do new things really  has stood the test of time. We see generative AI   as an even bigger step forward in terms of the  augmentation and enhancement of what it can do   for all of us in terms of giving us greater  tools and productivity to do new things.  I think the surprise, we did talk  about all this technology in that book,   and then our next book that my co-author and I  wrote – Jim Wilson is my co-author – which was   called Radically Human. That was the second book. The pace of the advance is what surprised us   more so than the capability. We were anticipating  that some of these capabilities would come along.  But the pace of development of the  foundation models, the rapid growth,   the size and complexity of the parameters, and the  weightings and everything, and the breakthroughs   that came about with that, were probably the  biggest surprise, Qu, in terms of what we saw.  Then one last thing on that. When you talk about  the timing and how fast everything is coming  

together, it's fascinating to think that even  open AI's ChatGPT is still relatively nine to   ten months old (as we stand today). Yes.  When we fast-forward to just yesterday, Elon  Musk announced XAI, which is another fascinating   AI company. As a business leader and executive,  how should I think about AI? It's happening fast,   but does that mean I take the "move  fast and break things" approach,   or should I wait and see where things settle? On the flip side of that, an organization   might be behind. How should I think of that? Our belief is that generative AI is a participant   sport. You have to jump in and start using it,  experience it, and do some experimentation. We're  

encouraging companies to do that, and that's the  approach we're taking in our own organization.  It's very early with the models. You just  highlighted that with how young the GPT and   ChatGPT models are. A lot of companies have not  reached GA (general availability) status of their   models and products, so it's early evolving. Elon's company was announced recently, and   there are new companies sprouting up continuously.  And so, I think the key for companies is, first,  

look across your business and decide where it's  applicable. Second, pick some use cases where you   can jump in and experiment with the technology  and manage some of the complexity and risk. Then   third, develop the foundational capabilities  that you need to then scale it faster.  Those capabilities include technology capabilities  like understanding the models: the problems   engineering, the pretraining, and other things  that you might need to do and how to integrate   these models back into your business. As well as  the business skills of understanding how and where   you apply it. How you develop a business case for  it. How much does it cost to apply these models?  Those three steps—looking across the landscape,  experimenting, and laying the foundation—are   what we're helping a lot of companies do today. Be sure to subscribe to our newsletter. Subscribe  

to our YouTube channel. Check out Paul, you're describing this kind of open field   of innovation that's going to be happening. But  everything around generative AI right now seems   so ambiguous. The technology is changing. The  implications for business are apparently amazing   but unclear. And so, how should business  leaders navigate this intense ambiguity? 

I think generative AI is just a new ingredient  into the mix. We've been talking for a while; I've   been talking for a while about the exponential  advance in technology that we're living in. We've   been talking for a while about how organizations  need to develop the ability to innovate and   recognize and adapt technology faster. The three key technologies that I think   will define a company's success in the  next several years and decade are cloud,   artificial intelligence, and the metaverse.  Those are the three themes, and I can talk more. 

We're talking about AI today. I'm happy to  go into others, other directions, as well.  As you look at the AI piece of it, those things  build on each other. To be successful with AI,   we're finding, and companies are finding,  they need to get to the cloud. Those that   have an advanced foundation in the cloud  are better prepared how to utilize AI. 

Most of these models run in the cloud, and  you need to have your data foundation in   place. Have the data to drive the AI model  successfully. A lot of organizations have   struggled with this over the years. We did a recent survey and only 5% to   10% of companies really have maturity  in terms of how they manage their data   and the corresponding AI capabilities.  That means 90% have a long way to go.  You need to start with the cloud foundation  and what you're looking at. You need to look   at your data, the governance around your data  and your metadata, and how you pull all that   together so that you can support AI in the  right way. And then it's the AI capability  

and skills that you build on top of that. It's a journey that we're on, and it's going   to continue. Generative AI is amazing, but  it's not the last big breakthrough and it's   probably not the biggest breakthrough we'll see in  technology as this exponential advance continues.  This is kind of the muscle, so to speak, that  organizations need to develop to continuously   anticipate and have the flexibility in their  systems, their architecture, their business   and their business processes, and their talent  to continue to adapt as technology advances. 

From your perspective, AI is essentially  another (in a chain of technologies)   that's not necessarily all that  different from what's come before.  What's different about AI... It is the latest in  a chain, and these things all build on each other.   It's this combinatorial effect of the technologies  coming together that really creates the power.  But what's different about AI is it allows us  to create more human-like capabilities. I can   communicate with large language models using  natural language, using voice interaction,   et cetera. I can get output that is  easier for me to interpret. That's the   powerful breakthrough with generative AI. What I advocate is the more human-like  

the technology, the more powerful and the more  exciting it is for us. We shouldn't view it as a   threat (as technology acquires this capability).  It allows us to really leverage the technology   and give us superpowers (as we talked about in  our book) around giving us new capabilities.  For example, I can be a customer service  rep and, rather than just what I know in   my memory and from my experience, I can have at my  fingertips every aspect of every technical manual   on the product that I'm answering questions on  brought to the forefront and prioritized so that   I can ask the questions right way. This is the  type of power that the technology has given us.  Just to g at that a little further, the real  impact of AI – while cloud probably changed   technology a lot and how we build technology  and support technology – AI is changing work   and the way we work because of this capability.  One of the research studies we did recently showed   that 40% of working hours across companies  globally are impacted by generative AI, 40%. 

That doesn't mean 40% of jobs go away.  Far from it. We actually see it enhancing   jobs and enhancing productivity and  capabilities that people have in many   ways. I'm happy to go into that in more detail. QuHarrison mentioned that your book was called   Human + Machine, and we have a really interesting  question from LinkedIn. This is from Milena Z. She   says, "How would you describe the significance of  incorporating human values into the development   of generative AI technology? It's incredibly important.   If you don't have, in your organization,  a really strong, responsible AI program,   you're simply being irresponsible. At  the core of responsible AI is accounting  

for human values in the way that you do it. Responsible AI, in our view, is about things   like the accuracy and coming up with the right  answers, avoiding the hallucination. It's about   the ethical issues that you need to think about  in terms of how you're applying the AI. It's about   bias and ensuring that there are fair outcomes and  fair use of the technology and, in certain cases,   the transparency and explainability  that you need around the technology.  We encourage every organization using AI,  especially with the advance of generative   AI – we've been talking about this for six  years, but especially with generative AI – you   need a responsible AI program in place. If you  can't inventory every use of AI in your company   and understand the risk of it and know how you're  mitigating those risks, then you're simply going   to get yourself in trouble with improper uses of  AI, and that's the way we think about responsible   AI. It's not just mushy values and principles.  It's execution, operations, and compliance  

in terms of how you're applying the technology. Paul, it's a great point. But the question I have   is the theoretical version of that and the actual  application of that oftentimes looks entirely   different. For example, if I'm in a company. Let's say I'm experimenting with generative AI and   it's just in our R&D department. Then we quickly  realize that this could actually have some scale.   We just apply it to a whole sector, another sector  of our company, or maybe we apply it to the whole   company. At what point do I actually stop and  say, "Okay. There's a legal component here"?  Oftentimes, when we point to that direction,  that's the big debate in AI today,   even at the congressional level, is what  do we do? How do we regulate this stuff?  If I stop now, aren't I hindering my  innovation? And if I'm in charge of   innovation and acceleration of technological  development within the company, I'm caught in   the Catch-22, if you understand what I'm saying. I am not one of those that support stopping,  

banning, or pausing on the technology. I think  it's about putting in place the right framework   and the right guidelines so that you know what  you're doing and can evaluate the risk of it.  I would say, not just at the end but every step  along the way, and before you even get started,   you should do an assessment of the risk. There  are a lot of guidelines and ways you can do that. 

The EU is going through the stages of approval  on the AI Act. They identify high-risk,   different risk categories of AI. Does your team understand those,   and are you assessing for any application of  AI? What risk category are you fitting into?   And then how do you mitigate that or deal  with that to make sure you're handling that?  That's just one example in respect to EU.  There's also the Whitehouse guidelines.   There's NIST and other things that are out  there. And there'll be more coming because  

of the interest in setting some guardrails  around this, which I think is a good thing.  But I think the teams need to be trained  and organizations need to have tools in   place so that you are assessing the use  of AI and, again, understand the risk   of it and make decisions accordingly. There are things we won't do and things   we've decided not to do, applications that we've  not pursued because of the risk profile of them,   or we felt it was not aligned with  the right values. That's an important   consideration to build into your process. It can't just be after the fact. It's got to be  

as you're considering use cases and starting out. AI brings out either the best in people or the   worst in people. The latter component, when it  brings out the worst in people, traditionally what   I'm seeing is people will try to hinder the AI's  abilities or slow it down in fear of losing their   job or seeing other calamities ensue within their  industry. One of the questions I have for you is,   how do we get better at communicating AI? Technology is neutral. Technology isn't  

good or bad. AI, generative  AI, fits into that description.  Generative AI isn't good or bad. It's  exactly as you said, Qu. It's how you use it.  It can be used for bad purposes. It can  be used to spread misinformation at scale,  

deep fakes, and all sorts of things. But  that's people using the technology badly.  I think that's what some of the communication  we need to do around generative AI is that the   thing we really need to be looking out for  and preventing is bad uses of AI and people   using AI in bad ways. We need to educate the  general population on what that means so that   they can recognize and understand if something  has been propagated and generated artificially   at scale using generative AI for some  illicit purpose, whatever it might be.  I think that there is a broad education that  needs to happen there. We're doing a lot of work  

on that. We're working with a lot of different  organizations on that, governments and other   bodies, to look at how we can better educate the  general population as well as business leaders,   technologists, and decision-makers around  using the technology in the right way.  I think that's an ongoing effort  that we all need to work together on.  We have a bunch of questions that are stacking  up on LinkedIn and Twitter. I have to say  

you guys in the audience, you are so  intelligent, so smart and sophisticated,   and your questions are absolutely great. Our next question comes from Florin Rotar.   He is the chief technology officer at Avanade.  I have to say that I did a video with Florin   years and years and years and years ago in  Seattle. Florin, it's great to see you pop up. 

Here is Florin's question, and I think it gets  right to the heart of some of the key issues.   He says, "How will generative AI change the future  of work? Can it also play a role to enable people   to realize their full potential, to thrive and to  grow, not just to drive productivity? Will it blur   the lines between white collar and blue collar?" I'll just add to that. To me, this question is   also getting to the point that QuHarrison just  raised, which is, generative AI brings out the   best and brings out the worst (in people). We talked, in Human + Machine,   about the idea of no-collar jobs,  and exactly what Florin highlights,   eliminating this distinction between blue-collar  and white-collar, as you look at it. Think about  

a hands-on service technician. Think about a  plumber or an electrician that now has access   to large language models that give them tremendous  amounts of additional information and potential.  It can give them tools to run their business  more effectively. Maybe they can be a service   provider to others in their profession  rather than just being the specialist   at the physical trade that they have. I think  that's the blurring capability that AI allows. 

Think about a small business (or any part of a  larger business) that wants to go international   overnight. They can start communicating in  dozens of languages seamlessly in expanding   their business. It's these superpowers that  are enabled that give people more capability,   and that leads to a lot of new  entrepreneurial activity and ideas.  Think of what GoDaddy did to the Internet in  creating a generation of entrepreneurs in a   lot of different ways, or eBay marketplace,  and such. We're going to see that to the  

next exponential multiple with generative AI,  creating all these new possibilities of what   people can do. That's what we see happening there. To get more specific around it, we see the new   opportunities for jobs and the way generative  AI impacts that fall into five categories.   The first is advising. This is advisors,  assistants, or copilots to help people  

do their jobs more effectively. For example, a large European service   organization that we're working with where we're  using generative AI in their customer service   organization to allow them to answer questions  with a lot more accuracy and quality because,   as I mentioned earlier, they can pull tremendous  amounts of technical information together to   answer customers' questions better, faster, with  higher quality, and they can cross-sell more   effectively because they get the ideas and prompts  and support on how to cross-sell. That's advising.  Creating is another whole category, is a second  whole category, another category. A good example   here is the work we're doing in the pharmaceutical  industry where we're able to (in the drug   discovery process and clinical trials process)  create some of the regulatory and compliance   documents they need to create so that then gets  reviewed at the final stage by humans in the loop,   avoiding all the rote work that a person would  normally do, and allowing them to apply their   judgment and expertise in the final product.  That's creative (in addition to applying it   in marketing and other areas that I could  talk about, which is super interesting right   now). That's the creating side of it. There's automating where you can use   generative AI to automate some of the  transaction processing. An example here  

is a multinational bank. We're using generative  AI in their back office processing to read and   correlate tens of thousands of emails that come  in with transaction activity. Normally, people   need to sort through all this to reconcile and  do their post-trade processing more effectively. 

Again, you can do this with other technology.  You can do it with generative AI. You can   make people's jobs more productive and  effective and take out some of the drudgery.  The fourth category is protecting, which I think  is super interesting. An example here is we're   working with a large energy company on a safety  application so that workers in real-time can get   all the information on what's happening.  Real-time conditions, weather conditions,  

and other things in a complex, say, refinery, and  then combine that with all the information they   need to know from safety procedures in manuals and  regulations and such that they can operate in a   more safe manner in real-time. Again, couldn't  put all this together before generative AI.  Then the final use case we're seeing a lot is  in technology itself, using AI and software   development in technology development. I'm sure we'll find more examples as   we go. Those are five that are kind of  standing out right now, just to drill into   some of the ways that it's transforming  work (in response to Florin's question).  We've got another question in from Twitter  from Chris Peterson. The question is, "One   of the opportunities mentioned in Human + Machine  was the AI explainer role. Is that even possible  

for something as complex as GPT4 with billions of  parameters and almost unlimited training data?"  In some industries, in some problems,  if you can't explain it, you can't do   it. That's part of that screening that I  talked about earlier with responsible AI.  If you have kind of a regulatory or ethical or  business need to explain exactly how something   is happening, you need to use the right  type of approach (where you can do that),   and you can't do that with (to your  point) some of the models that are there.  There's a lot of advance happening in  explainability. There are ways to create   the models to understand how they're processing. There are areas like GAN (generative adversarial   network) that we can use in different ways to  get some insight into how models are working and   such. So, there are a lot of different advances  there, and there are new fields, in addition. 

New fields like prompt engineering are  cropping up because of generative AI.   We're also seeing demands in the market for  explainability engineers or explainability   specialists who can bring that understanding in  to help understand those kinds of conditions.  The other thing that's sometimes  important is that, in some applications,   you don't necessarily need to explain exactly  how you got the answer. You need to provide the   transparency of what information you're using,  what data you're using, and the process itself.  You need to differentiate where you really need  to explain exactly all the math you did and how   you did it, so to speak, and where you just need  to provide transparency into how you're doing   it and show that you're using information such  in the right way. Distinguishing that can help   organizations unlock some of the potential, too. We have another question from Wayne Anderson. You  

can see we love taking questions from the  audience. Again, the audience is amazing.  This is from Wayne Anderson on Twitter. Wayne  also has a question coming up on LinkedIn,   so he's sort of a multi-tenanted— Multi-platform. 

–multi-faceted social media happening here. Wayne  says, "What is the litmus test? Is there one,   a question, set of questions, that you use  to quickly evaluate a client's place on the   operational maturity journey for AI and ML?" We have a maturity framework we use to assess   for ourselves as for our clients. There are steps  of maturity that you go through in assessing it.  There's assessing talent and where you  are with the talent and the expertise   that you have in the organization. That's  about the technology talent as well as the  

skills you have in the business and the kind  of training programs that you have around that.  There's assessing the data readiness for it  in terms of (as we talked about earlier) your   data maturity and the maturity of your platforms,  data platforms, to support what you need to do.  There's then the maturity of how you need  to use the models and your sophistication   around that. That depends on the strategy that  you have. Is your strategy to use proprietary,   pre-trained, publicly available models, or is  your strategy to do some of your own pre-training   or customization using your own data? That  requires far different operational skills and,   therefore, you need to evaluate  where you are on that spectrum. 

Then there's the operational skills around  it. So, how do you put the AI in place,   and how do you monitor it on an  ongoing basis for the right outcomes?  Then finally, the responsible AI dimension of it. ) Those are kind of the dimensions. There are more   underneath that. But there's a process that we  use to go through it, and I think that's every   organization having an understanding of that  and having a way to evaluate their maturity is   important to know how you're making progress. Wayne actually did ask another question on   LinkedIn. I'm looking at it right here. He talks about the security and the risk   of AI is not something that is entirely a  technical solution. A lot of it is in the  

humans and innovation/development processes.  What formal steps do you need to be in order   for that innovation to provide the  kind of guiderails and talking points   on the future of machine learning projects? The way I interpret that is we've got a lot   of groups working together. How do you make  sure they're all working and their energies   are going the right way, the right direction? We think the right approach to use is a center   of excellence kind of approach given the  state of the technology where you create   a center of excellence. You have to centralize  your organization that has those capabilities   in it. That's what we've done for ourselves  and we're helping a lot of our clients do.  In fact, we have something called a COE in  a box that we're using to help clients set   up these kinds of capabilities. It requires the  technology capability, the business capability,  

the legal teams and capability (legal  and commercial), and talent (the talent,   HR kind of capability around it). You need to bring all that together,   a center of excellence where you  can have that capability assembled.   Representatives from all those different  groups in your organization is important.  You can federate some of the experimentation then.  But it's really important to bring it together. 

Security is a great angle. I don't know if  that was the primary thrust of that question,   but there are a lot of implications  on security from generative AI,   both in terms of new security challenges as well  as consideration around data privacy, grounding of   models, use of sovereign data. Depending on the  jurisdictions you're operating in, those become   really critical considerations for companies. Having this built into kind of a center   of excellence that you know that you're  channeling this in the right way in companies,   I think, is critical for the stage of  development that we're at right now.  Paul, given your purview and some  of your thesis around the future,   one of the things that I'm wondering in your realm  is, when I look at technologies like the cloud,   an enterprise corporation is probably  best suited for that realization today,   right? Personal cloud computing, it exists,  and I think the strongest use case of that is   probably video games today. But beyond that, it  doesn't make that much sense for an individual  

person or even a small startup to endeavor on  a very complex cloud implementation. However,   I think that might differ given some of your  comments you just specified when it comes to AI.  On the AI front, one of the things that  we're seeing is corporations that have a   lot of technical debt or have a lot of data that  hasn't been digitized or have very complex teams   and org charts. They're not well suited  because it's going to take them some time   to get all these things in progress, in place. Now on the flipside, they have the most data,   so they'll probably have some of the stronger  AI models. But the question that I have here is,   would it make more sense for a startup or even  an organization to think about creating an   internal startup and then going after it? That's what Google did with Deep Mine,   and we just saw some of the news related to  Deep Mine where they're bringing in Denise to   lead their actual AI practices at Google. There are countless examples where this  

is also true in the AI industry. Is that  the right approach or do you think that   that ship has sailed a long time ago? For some companies. There's an example,   a media organization we're working with that sees  an opportunity to really create a whole new part   of their business using generative AI. They can use generative AI to create a   way to generate coverage for things they couldn't  cover before. I can't get too specific about it.  In that case, that's maybe more of a  startup. You actually are using generative   AI to branch out in a new direction. We think a lot of the generative AI  

potential is going to be changing the  core of how you work as a company. It's   going to transform the way work is done. That's that phase we used, "Reimagining   work." That's what this is about, which means  I think you do need to have a lot of capability   at the heart of your organization looking  at how you do and drive the transformation.  I think it could be a mix for different types  of use cases. A company may spin out or have a  

separate project to pursue some initiatives  they're doing. But I think this gets to the   core of how companies are operating, which  is why companies need to embrace it broadly.  But another point that you're mentioning  is I do think that generative AI offers   a lot of potential for new startups and  small companies because they can access   tremendous capability to build new businesses in  addition to the power it gives big businesses.  I heard people ask, "Are the big companies  only going to get stronger with this?" or   "Are the new startups, new companies going to  win out?" I think it's really a mix here that   we'll see going forward because of the power of  the models and the power for new organizations   to leverage them as well as the power that  larger organizations have to move faster. 

Paul, let's shift gears a little  bit and talk about investment,   technology investment. AI is changing so rapidly.  The capabilities are changing. The models are   changing. The implications for the enterprise,  and for society at large, remain very unclear.   Given this ambiguity, how do you recommend  that organizations should be investing?  I will mention that Accenture recently announced a  $3 billion investment in this. Obviously, this is  

something that you're giving a lot of thought to. As you said, we announced a $3 billion (billion   with a B) – we don't do that too often - $3  billion investment in data and artificial   intelligence. There's a good part of that is for  generative AI, but it's across data and artificial   intelligence, so we're doubling our workforce. We have 40,000 people that work in data and   AI today. We do a lot of work in that area.  We're going to double that over three years.  We're developing a new tool called AI  Navigator for Enterprise to help companies   apply AI more quickly, including generative  AI. The tool itself uses generative AI to   help companies understand the roadmap they  need to follow and, industry-by-industry,   how they can drive value from AI. We're creating a center for advanced  

AI where we're looking not just at generative AI  but the next breakthroughs that will come as well.  Yeah, we're excited about it. We're putting  a lot of money and focus on it because we   do believe this is transformational  for business and this wave will build   faster than cloud and faster than some of the  other technology waves that we've seen before.  Yeah, a big focus, and we see companies  doing the same. We did a survey recently,   and 97% of executives that we surveyed—this  is just a couple of weeks ago—believe this is   going to be strategic for their companies, and  it's going to change their business or industry.   Ninety-seven percent, that's basically everybody. Over 50% believe it's game-changing. Not just  

some change, but game-changing  for their industry or company.  About 46% are going to invest a significant part  of their budget in generative AI in the next 2   years. This is a fast build, and maybe some of  this is companies getting a little over-excited,   but we believe that that pattern will hold  and companies will move and invest in this   technology more quickly than we've  seen other ways of technology built.  But what about the risk associated  with investing in something where   the end trajectory is so unclear? You need to look at the horizon. I   think there are a lot of things that are clear. I think the key thing is to look at this from   two dimensions: business case dimension and  the responsible AI dimension, which helps you   balance the risk. The business case helps  you look at the value. The responsible AI  

helps you look at applying the human  values and the right risk profile.  If you take those two lenses, I think you  can find the intersection of the right   things you can start on now with no regrets.  Obviously, you have to make sure that the use   case you look at can be supported with  the technology that's available today,   which is moving super-fast. I think, Michael,  you can identify no regrets things to do.  We believe, in the near term, this is going  to be human-in-the-loop types of solutions for   the most part. It's going to be solutions  that bring in tremendous new capabilities  

for people. It's going to be new, exciting  capabilities for consumers to use more directly.  In one case, a retailer we're working with  is using generative AI to create all sorts   of new product configuration capability  for their customers. It's going to create   new capability for employees, et cetera. This is all stuff that's doable today,   I think, with no regrets, without really worrying  too much about the risk. You can apply the right  

principles to do it in a responsible way. From an industry-specific standpoint,   it seems like each industry is dealing with  AI at its own speed. The two that I want to   bring up right now that have had probably,  I would say, some of the most impact is,   one, education, and two, the legal sector. The funny thing about it is they dealt with   this regulation in entirely different ways.  In the education sector, everything is pretty   much a chaotic mess. You have schools banning  things, turning things off, then re-enabling.  

We could have a whole show on this but, on  the legal side, you've got—which surprises   me the most as a technologist—lawyers  really embracing this technology.  There's obviously a little  resentment, but there are legal LLMs,   and there's a lot of adoption as to how you can  integrate it and adopt and make your law firm   or your practice move faster. I would have never  predicted that in 100 years, but it's happening.  Now, on the flip side, Lisbeth Shaw from Twitter  has this really good question where a lot of   organizations and individuals have begun  using generative AI for work without any   AI governance in place. She's wondering  how you can apply governance once the   horses are out of the barn and racing. The reason why I brought up the points   earlier is education. That whole sector  is dealing with this whole dilemma right  

now. I'm curious on your take just because  you're seeing it on the enterprise side where,   if I input an email or contents of a document,  there is a true risk there whether it be IP or   trade secrets, whereas with school, if I put my  quiz questions and test questions in the program,   it really only impacts me and will have a lasting  impact on the knowledge that I retain and gain.  We're seeing broad adoption across industries,  unlike any other technology I've seen which   had very specific, and everything had specific  industry patterns. Client-server, ERP, mobility,   cloud, SaaS, had very specific industry  patterns. Generative AI is super-broad in   terms of the industry adoption we're seeing and  the industry potential use cases we're seeing.  The two you mentioned are super  interesting, Qu. Education, I think,  

will be literally transformed through generative  AI. It enables truly personalized learning in   ways that are significantly different than  our current educational system. It'll take   a while for that to work through but, yes,  it's going to be pervasive and powerful. 

Legal, I agree with you. The interesting  thing about the legal profession is it   can help paralegals work more effectively  and do higher-level work, and it can allow   experienced lawyers to leverage themselves more  effectively in terms of the work they get done.   So, we're seeing it being adopted across the  different types of work in the legal industry   or industry profession from that perspective. But I think to the horse out of the barn question,  

you can still apply responsible AI.  You can go back through and do it.  It's a matter of being systematic and rigorous.  It's about having C-suite and CEO support.  We report on responsible AI to our board. It's  part of our formal compliance responsibility  

that we do. And we encourage  organizations to do the same.  If you already have AI out there, and most  organizations do, and most organizations   don't have enough responsible AI in place,  we believe it's time to do that. Inventory   the AI. Know where you're using it. Understand the risk level. Know the   mediation techniques and tools and have them  at your disposal. Know if you've mediated the  

risks. You have to go back retroactively and  do that if you haven't done it so that you   know what your baseline is as you start to  apply more AI and generative AI going forward.  Given the impact of AI, we know that it is  profound and will be profound. Where is this   going and, more importantly, how should businesses  position themselves to capitalize on this obvious   sea-change that's kind of erupting all around us? I think the simple answer is you need to think   big, start small, and scale fast. The think big is think about what the   real potential is and where this could take your  organization. Where are the big threats and the   big opportunities? That's thinking big. Start small is the experiment with the human  

in the loop and the no regrets use cases. Get some  experience. Understand the model. Select the right   partners (models and such) and do something. Get ready to scale fast. This is the centers   of excellence, the operational maturity  that one of the good questions came in on,   and other capabilities, and the talent  that you built around it to scale fast.  "Think big, start small, scale  fast" is the advice I'd give.  Sci-fi has shown us what the future looks  like. We see some of the gadgets and gizmos  

that are real-life objects from Star Trek. We  see some of the unforeseen and uncomfortable   futures from Black Mirror start to arise. One of the things that I'm wondering your take is,   I mean, you wrote the book Human + Machine,  and then you've got another one since then.   I'm guessing you've been thinking about this  whole concept of transhumanism and merging the   brain-computer interfaces that Elon talks about  with some of these AI models. How near do you   think that is, or do you think that that is  still fodder for science fiction novelists?  First of all, I'm a massive  fan of science fiction,   and I believe most science fiction eventually  becomes real. It's a matter of the timeline.  If you want to read about where technology  is going, you pick up somebody like Neal   Stephenson and read his books where he  coined the term metaverse among other things,   and his book Fall previewed where we are  with technology right now (a number of   years ago) really well. Science fiction can be  incredibly illuminating into where we're going. 

In terms of transhumanism, I'm not  a real expert per se in that field,   but I talk to a lot of friends and colleagues  who are. I believe it's quite far away.  Think about how blown away we are by large  language models today and ChatGPT and   everything. There is no intelligence inherent  in these models. These are statistical models.  People ask me how intelligent these models  are. The models have no intelligence. The   models are a bunch of data with technology  that can statistically create results from   them. There is no inherent knowledge. Now, some of the breakthroughs we're   looking for in AI, the next generation of things  like common sense AI, the way knowledge graphs   come in and can be combined with generative AI,  that starts to create systems that have more   intelligence inherent in the models, along with  the generative capability. I think that's where   you see some interesting advances. But truly getting to the human and  

surpassing human level, I think we're  quite far away from it. We're multiple   breakthroughs away, I believe, from seeing that. I think that discussion distracts us a little bit   from what we need to do today, which is some of  the great questions that listeners have asked   about human values and ethics. Let's prevent  people from using today's technology in bad  

ways and avoid getting a little bit too distracted  by the things that are pretty far down the road.  This is from Mike Prest. He's a chief  information officer on LinkedIn. He says,   "As a business leader managing the risks of AI,  what advice can you offer on sharing information   to become good stewards of the technology and  dispel some of the dystopian conversations   about generative AI?" Very quickly, please. I think we should share more. On that front,   I'm happy to connect with  anybody and share some ideas.  There are various forums out there where there's  a lot of this sharing happening (both in business   communities and different technology forums). I  think that's how we'll all get better. It's at the  

early stages, and I have a lot of forums that I'm  running with some of my peers and colleagues and   other companies to share a lot because we're all  learning together in this fast-moving technology.  We have another question from Twitter, another  really good one. Again, really quickly,   please. This is from James McGovern who says,  "With Microsoft and Oracle holding layoffs,   the talent for enterprise architecture and sales  professionals must be huge. Who is hiring?" 

Enterprise architecture: As much as you need  generative AI skills, enterprise architecture is   immensely important. Generative AI (along with the  metaverse capabilities, which we didn't talk about   in this call) creates really a rethink of your  enterprise architecture and what you need to do.   So, those skills, I think, are in tremendous  demand as we look at this going forward.  I think a lot of companies are looking at hiring  the right talent to build this out. And I think   enterprise architects, in particular, have  been a shortage in the industry for a while   and are even more in demand as we go forward  with every new technology like generative AI.  Paul, let's shift gears here. You're an  avid sailor. I've known you for many years,  

and I see you sailing. Tell us why. Tell us about  your sailing. Why do you like to sail so much?  I've sailed my whole life, so it's  something that's been a lifelong passion.   I love the experience of it. When you're out  on the water and you're seeing the sunset, you   have a nice breeze behind you, and you're powered  only by the wind, sailing along, and can hear the   bubbling under the keel of your boat as you're  moving through the water at a nice pace, there's   not a better feeling in the world than that. There's a challenge aspect of it, which is  

optimizing. How do you go a little  faster? How do you get the sails   tuned a little bit better? And I love  the intellectual challenge of that.  There's a learning aspect. I learn something. I've  been sailing my whole life. I learn something new  

either by making a mistake or just encountering  something every single time I'm on the boat,   and it's a continual learning experience. Finally, I'd just say it's my happy   place. It's the one place where I really  don't think about anything else because,   from a safety perspective and focusing on what  I'm doing on the boat and everything else,   when I'm on my boat, that's where I am and that's  where my whole focus and my mind is on my boat and   the guests and passengers that I have on it. As an author, I'm sure some of your pastime   includes reading. What books are you reading  these days, and what's keeping you sane?  One of my favorite authors and heroes is  Neal Stephenson who wrote so many great   science fiction books, so I'd put him out there. A great book that I read recently is Cloud Atlas,  

which is a fantastic story that gets into  some of the topics that we talked about. It's   a prize-winning novel that covers everything from  the fall of the Ottoman Empire to space travel in   the future (through a series of parallel  stories). It's a very interesting read.  There's a book called Reality+,  which I'd recommend to anybody,   anyone that's interested in, first of all, the  transhuman topic you mentioned, the metaverse,   or related topics. Reality+ is by a philosopher  from NUI who is exploring the question of are we   living in a real-world or a simulation, and how  would you know the difference between the two?   It's a fascinating book and super well-written. I read a lot, and those give you a sense of the  

realm from fiction to science fiction  to philosophy as well as technology.  You're the senior person for technology  at Accenture, which employs about 740,000   people. Just that number in and of itself  is almost incomprehensible. How do you   spread yourself over 740,000 people and  manage the pressure and the expectations?  It's an amazing privilege to have a role  like this. Our mission is to deliver on  

the promise of technology and human  ingenuity. The human ingenuity that   we have in those 740,000 people is just amazing. What I like most about my job is the ability to   learn from 740,000 people. I don't talk to each  of them individually, but the work that we do  

for clients, the innovative ideas they come up  with is just super inspiring, the projects we   do in terms of improving communities and  society through some of the work we do.  It's really a privilege to do it. I'm  just honored to have the role and to   represent the amazing group of people that we  have and the amazing leadership that we have.  It is a big company. It's a lot  of people. But it's a lot of  

small communities that come together with a  common culture is the way to think about it.  We have the system that we know how to  hire people in volume if we need to. We   know how to build community and build culture  in our organization in a lot of different ways.  As you scale up and get bigger, some things  aren't that much harder to do at a bigger   scale and upscaling very well as you grow. That's  what I've found as we've grown the organization. 

It's a lot of fun and, again, it's just a  privilege to be in an organization like this   and have the role that I have. What's the hardest part?  I don't know all the 740,000 names, but  I'm working my way through as best I can.  Hey, Paul, a question for you regarding just  being a techie. What's your favorite device?  Probably apps that I use. One of the devices I'm  really getting a kick out of is my Oura Ring.   Not to do any marketing for a specific  product, but it's a simple device. 

The ring is connected to the app on the phone. And  I'm finding it's really helping me understand some   patterns and how I can be a little healthier  and happier and get better sleep and such.  You can track. I can track  and correlate my heart rate,  

my oxygenation, my breathing  patterns, all sorts of things,   compared to my sleep activity, compared to my  sleep cycle, compared to my activity cycle.  We're data-driven, and if you get better data,  you can improve patterns and such. That's one   of the things I'm playing around with right  now that I'm getting a lot of value out of.  One of the things that's interesting about the  Oura Ring is it represents the whole quantified   self-movement. Right.  You now have your own personal database of data  that you can do whatever you want with. Are you  

going to build anything using your health  data or is it just a personal experience?  I don't know, but I'm on that exact journey  you mentioned. I'm starting with now the   personal biome, understanding your biome  more using the self-diagnostics, which has   another big impact on health and wellness. Yeah, I've been trying to get more and more   data-driven and understand what makes me  work and what makes me healthy or not. Yeah,  

that is something I'm going to continue doing. It's funny because that's the big data that comes   off of your body, and then you could take that,  what works for you, and implement that at the   enterprise at scale. I see what you're doing. [Laughter] Exactly.  [Laughter] Okay. With that,   we are out of time. A huge thank you to Paul  Daugherty. He is the chief executive for Accenture   Technology. Paul, thank you for coming back again  to CXOTalk. We really, really do appreciate it.  It was a pleasure, Michael, and it's great to  do this with Qu as well. Thanks to you both  

and to the audience. Those were amazing  questions. I wish I could be there and   ask the audience a lot of questions as well,  but it's been a great experience. Thank you.  QuHarrison, it's great to see you. Thank  you for being such a great co-host.   That was a lot of fun, wasn't it, Qu? Indeed, man. Thank you for having me. 

Everybody, thank you for watching. And as  Paul said, you guys are an amazing audience.  Before you go, be sure to  subscribe to our newsletter.   Subscribe to our YouTube channel. Check out, and we will see you again next time.   We have amazing, really great shows coming up. Have a great day, everybody. Bye-bye.

2023-07-27 11:37

Show Video

Other news