DRUID Talks Season 2 Ep#5: The Role of Conversational AI in Navigating Technological Disruption

Join host Kieran Gilmurray as he welcomes Sultan Mahmood, a seasoned tech visionary and digital transformation expert, to DRUID Talks. With nearly two decades of experience in technology consulting and a deep involvement in cutting-edge innovations like VR, Web3, the Metaverse, and Conversational AI, Sultan is a go-to authority for businesses and individuals looking to thrive in the digital era.

Kieran Gilmurray:

Welcome everyone. Sultan is a hugely experienced senior executive. It helps people and businesses succeed in the digital age through a radical transformation of operating model, process change and innovative use of technology. He has worked in tech consulting for almost 2 decades, SAT on boards of tech startups, and explores the latest VR, Web 3, Metaverse, blockchain, and conversational AI technologies on a daily basis. If you want to understand tech and business, then Sultan needs to be on your go-to list.
Sultan, welcome, and I am delighted to have you here today, Sir.

Sultan Mahmood:

Thank you, Kieran; great to be here, and thanks for inviting me.

Kieran Gilmurray:

Sultan, let's jump into the questions as we've got a lot to talk about over the next period of time. It's fair to say the speed of change in business as driven by technology over the past 12 months has been massive.

How do companies start to make sense of all this change? To know where to invest, to train their people to design their business strategies, and everything else in between?

Sultan Mahmood:

Yeah, that's a very interesting question, and I get asked that question by CX OS and board members quite often because, yeah, I'm sure a lot of people have kind of experienced their board member coming to them and saying what are we doing in the Gen. AI, what are we doing with this? Because they've obviously seen the latest article or the latest newsletter. And you know, that's a very difficult question that a lot of companies are wrestling with at the moment. So, you know, we know the pace of change has been huge over the last six months.

You know, especially with the likes of ChatGPT, Open AI, all those other things. And what we often talk to organizations is, well before you can work out where you should be and how you keep up with the piece of with the pace of change, you should really understand where our organizations are headed and what we've done as part of, you know, the work that we do is we've come up with a model which says, well, if you look at where organizations are headed and you look at the key technologies that are impacting organizations over the next few years, then what you will end up towards increasingly is what we call an IADVO, and that's an acronym. But it stands for Integrated Autonomous or distributed Virtual Organization. So IADVO is a model that we've come up with which says, how do you, where are organizations going to end up in four or five years time. And to be an IADVO organization, you're going to have to make use of technologies that are coming up and actually disrupting.

So, the ones that we're all familiar with are AI, machine learning, and conversational agents. Yeah, that's a key pillar of an IADVO organization. The other technologies that come into play that will, you know, impact and probably not as mature as where AI is at the moment because that's the hot topic of the day, technologies like Web 3. So, Web 3 includes blockchain and smart contracts. How do you create a distributed organization that functions in an autonomous way? And then, you know, the third technology that is going to impact and help us help organizations move to IADVO's is what we call Metaverse or, you know, virtual organizations.

And you know, I don't think Metaverse is dead, despite what we heard in the press a few weeks ago about, you know, Facebook and so on. Metaverse will reemerge. The fact that you know organizations will need a digital presence, and a digital presence that is as near as accurate as possible to a real-world presence will come back as well. So, when you combine Web 3, Metaverse, and AI, that's where you start building a fully distributed organization. And we've defined the four-box model, which is designed to help clients, you know, assess that. And that four-box model basically says if you're a board member, first you need to decide what sort of IADVO you want to be, and that's about deciding your goals.

An example might be.
So, for some companies, having a virtual presence would be really important. So, that would be part of your IADVODVO goals. For some companies, that might not be as important as making sure they have the best conversations with their clients. So, AI would obviously be higher up their agenda. The first step that we say to boards is, you know, to spend some time working out what your IDVO goals are. Once you've got the goals in place, then it's about, you know, developing your processes and operating models and working out which of the processes and which of your operations can be enabled by the technologies that underpin an IAIADVO. So you've got the technologies in the horizontal bars on the diagram on the slide there. So it's basically, you know, designing the operating modeling processes, and then you implement those technologies, and then you obviously go to the 4th stage, which is monitor and refines and continuously improve.

That's the overall model we suggest to organizations on how to, you know, move towards an IADVO.

Kieran Gilmurray:

I remember that sinking feeling well, by the way, when a board member had seen the latest technology in the press and decided that this may be the thing that would transform the organization. Well, Sultan, I'm going to come back to the model now and ask you an additional question. I'll pop it up on the screen as well so that people can see it here. There are four-box maturity models here. 

What are the technologies that help organizations transition from one level to another, and where do you see conversational AI fitting into that maturity model?

Sultan Mahmood:

So, the four-box model obviously is an overall approach, right? Within that, there are the different technologies, which you can see in the horizontal arrows at the bottom of the diagram. So I'll just briefly cover, let's say, conversational agents, right, conversational agents. We've defined 4 stages of maturity for conversational agents as a technology. So the first stage is all about, you know, FAQs and chatbots.

Lots of companies have played with FAQs and chatbots and have got, you know, instances of that in production and so forth. That's what we call stage one of conversational agents. The second stage is moving from an FAQ to a sort of chatbot to a more complicated conversational agent that's driven by AI and enabled by much more sophisticated NLP models that are coming into play. So, for example,  DRUID supports quite complex conversational agent conversations. So, I would say that's the second stage of making your conversations better.

The third stage of conversational agents in terms of maturity is what we call virtual assistants. And so you might say to me, what's the difference between a conversational agent and a virtual assistant. The difference is really the conversational agent can have great conversations with you. A virtual assistant can not only have conversations with you, but actually execute tasks for you, right? So, the third stage is all about implementing virtual assistants.

An example would be if you're having a conversation and you want to, let's say, book a holiday, then the conversational agent moves from a virtual assistant and actually assists you in booking that holiday and executing all the underlying transactions that need to be executed. The fourth stage of a conversational agent is basically having an enterprise platform that allows you to do all those stages consistently across the organization. So, at the enterprise platform level. Then you have the bunch of tools and technologies that allow you to implement FAQ where you need, implement chatbots where you need, implement conversational agents where you need, implement virtual assistants where you need, and bring all of that together under what we call an enterprise platform.

Kieran Gilmurray:

Because I do think, at times, people get confused between chatbots and conversational AI. So, I love the way you've described the two. Can a chatbot be a conversational AI platform, or is it the start of a conversational enterprise application platform or different sometimes?

Sultan Mahmood:

For me, it's the start of the conversational agent platform, and I love how DRUID has matured its platform. For example, it's gone from originally doing conversational agents more in the chatbot space to becoming a full conversational agent and now becoming an enterprise platform. So, that whole maturity in action can probably be seen with the DRUID Platform itself.

Kieran Gilmurray:

It's good to see. And I see a lot of companies starting with chatbots, FAQs, and then I've seen DRUID's knowledge base. Two very, very, very different things. One hell of a lot better than the other. One thing we cannot do, though, Sultan, is ignore ChatGPT and generative AI. Just like we can't ignore data security, privacy, regulation, and all these other things. 

How can companies make sure they obtain the best enterprise conversational experience using ChatGPT, again without compromising security or any of the good standards that they should put in place?

Sultan Mahmood:

Yeah, this is a topic a lot of companies are wrestling with. An example is: I was talking to the C-Suite of a major bank, and basically, they've said they've banned ChatGPT. They don't allow the use of ChatGPT in their organization, and the reason they've banned it is that it's uncontrolled and could potentially prevent a security risk. Obviously, there have been quite a few scare stories of people uploading stuff into ChatGPT, and then it kind of got into a domain where they didn't want that information to go. Now, ChatGPT is a great large language model, right? So what it does is it's been trained on millions of language documents, and it can predict the next sentences and work out what the next part of the conversation should be. And that is that has its part to play. So, if you want to have large conversational discussions with those who are human, then ChatGPT can be integrated into your environment. And to an extent, I think that's what DRUID has done. Yeah, they've integrated ChatGPT. It's not a competitor to DRUID.

DRUID is a solution that gives you the ability to have conversations based on a very large language model that's been trained. Now, does that solve all the use cases for an enterprise? No, it doesn't. And what else? What I think will emerge, and we see increasingly emerge, is that you move away from, you know, it can solve everything to, here's what it's good at, but here's a bunch of other AI models that are good at slightly different things.

So, for example, we're building models for an organization that wanted to be able to link user preferences to user products. To be able to train up on new products that come through, and if the user wants a particular product, it just, you know, brings that product out. Now, that's not a topic that's very well suited to necessarily a large language model; it might be suited to a much smaller type of model, which is using the word to VEC or one of these other models to analyze and produce and come up with similar products. So, for me, I think you know what you'll probably end up emerging is we're kind of doing some of that in IQEQ Digital, where you will end up with lots of different types of models that need to be within the enterprise platform that can be accessed through some sort of integration layer. And to an extent, that's a conversation layer. So, if you're having a conversation and you want to be able to reach into a general conversation, you might plug into ChatGPT.

If you want to be able to get preferences of what sort of products might be better for you and do a detailed mathematical analysis of those, you might reach out to a different type of model as well. So, I think that the enterprise platform becomes much more important for bringing all of that into play.

Kieran Gilmurray:

It feels a bit more fungible if that's a word, when you can do that. I love what you're saying there because I do see a couple of nuances in what you're describing there. I think boards sometimes see a technology solution. We had it with our RPA, where it's going to be the answer to everything, but it is a key. It's one key for many locks.

Again, I know the DRUID Platform integrates with Microsoft Azure Open AI, which actually allows you a tremendous amount of security privileges and practices. That means your data won't be exposed to some of the public as we have seen, and I'll not name some of the large companies in the world that did that, but interestingly, someone said to me recently, if you think your staff isn't using ChatGPT public, they probably are using it as well. See, you never quite know. You were describing a moment ago some of the technologies that you're involved in. I love every one of them and adore them. And I agree Metaverse is still here, that it's not gone away. 

Have you seen a technology that actually is ahead of its time? Or are you identifying something that's next on the list that will allow organizations to change faster, be better, grow quicker, and do all the things they want to do? What are you saying, Sir?

Sultan Mahmood:

All these technologies are in different stages of maturity, right? One of the things we specialize in is obviously looking into the future and seeing what is going to be the next wave of technologies. So, one of those technologies that probably is a bit early at this point in time is quantum computing. For example, quantum computing, you might have read in the press. A few days ago, Google announced some quantum computing processing that they did, which was, you know, the calculation it did was so fast it would have taken a normal computer thousands of years to do. So, if you look at quantum computing, it gives you an exponential increase in computing power, and its underpinning technology to others, to AI, VR virtual reality, and Web3, can underpin all of those.

It can actually present a threat to, for example, blockchain because, you know, with quantum computing, you can throw so much processing power at decryption that you can actually start decrypting standard encryption techniques as well. So, it's a technology that is one to watch. It isn't quite in commercial use yet, although there are some of the leading edge companies still playing with that. But I guess for me, it still moves you back towards, you know, achieving a proper IADVO. So if you've got AI empowered, you know, with quantum computing, you've got virtual reality empowered with quantum computing, you've got Web 3 empowered with a quantum computer, you will see an exponential increase in power. That means that some of the things that we probably can't imagine will appear in the next few years. I think there was a session I attended yesterday on generative AI with various specialists. And one of the speakers asked a question. He said: "What do you think will happen in five years' time?"

He said, "Look. If you had asked me in October last year what was going to happen in six months' time and everybody would be talking about ChatGPT, I would have thought I wouldn't have even dreamt of what actually happened." You know, the pace of change is so fast that you can't even predict more than a year and a half ahead of what's going to happen.

Kieran Gilmurray:

Oh my Goodness. I have to say, I adore being in this industry. I just wonder, with all this computing power, will it be a bit like Excel when you use 5% of it yourself, not 100%?

Sultan Mahmood:

I was going to say that one of these kinds of questions starts opening up the sort of self-questions: Is the computing power going to surpass what we as humans can do? One of the anecdotes I like to bring up with people is that if you look at how we've been trained in life, we've basically, as individuals, been promptly engineered.

Kieran Gilmurray:

That's true.

Sultan Mahmood:

We've been given prompts all our lives that give us the ability to do what we do, as you know, as professionals now, and the level of prompt engineering that we have had determines the kind of problems we can solve. So ultimately, you know there will be a role for humans because clearly, we're prompt-engineered with IQEQ, and we're prompt-engineered with life experiences. Now, how does that fit with a very powerful AI capability? I think we're seeing Copilots and all that kind of stuff emerge. I think there will be more and more of that kind of thing emerging where there will be a balance of what we end up doing versus what technology ends up doing for us. But ultimately, I'm a positive person, and I don't think technology will necessarily be able to replace everything we do.

Kieran Gilmurray:

I don't think it will ever have. I get what you were describing there with the copilot. It's the excitement that we can create exceptional technology with exceptional human prompt engineers to allow us even greater business outcomes. But see on that; so this is something I think businesses struggle with at times with all these new technologies: they're not quite sure how they get a return on their investments in technology because it can be hard sometimes to understand what is a what is AI or what is VR, Web3, Metaverse, conversational AI, and blockchain. When they don't know what those are and they need help understanding what they can do, then they're not sure where they should put their dollars to get a return. And as you mentioned a moment ago, things move so quickly that if I put my dollars in today, does it mean I've wasted my dollars?

How do companies pick the right combination of future technologies and build a business case around that to try and guarantee your return in 18 months, three years, or, dare we say, five years?

Sultan Mahmood:

Yeah. This boils down to even business strategy having to be accelerated. So the days of when you did a business strategy, which was a five-year business strategy, and then you visited every year, and at the end of the five years, you said, “Oh, we achieved our business strategy.” You start with the business strategy, and we help clients and boardrooms understand how these technologies impact your business strategy. As part of understanding your business strategy, what areas can you quickly get into? So, I think Move Fast is a recommendation that we would have for boards to understand the fuller implications of IADVO. You might not want to implement all of IADVO now but understand which areas will be critical for you in the next months rather than years and implement quickly in those areas. Still, keep in mind where the IADVO will take you, but implement quickly, which requires board education; it requires some fast thinking of how to move from a boardroom idea to a solution in weeks rather than months. And then, you know, how does that disrupt your business model? So, where will your business be able to come up with a different model of using those technologies? Where does it need to adapt itself and become, probably re-engineered or disrupted, or stop doing what you're doing, you know, start doing something different? So, I think the messages you start with the boardroom, you deliver and identify solutions fast that you want to say take forward; you learn from that.  

Some will work, some may not. And you quickly say, well, this doesn't work, and you go back to what are additional ideas we can bring into play as well? That whole approach is really important. And the timelines to do that approach have changed as well. So, in the old days, you could do some analysis or research, but now you need your best to get somebody who's already done that thinking and research and seek advice from them to quickly move from what we call ideations to solutions. 

Kieran Gilmurray:

So we must find someone who has played on the bleeding edge before us. It's interesting because as you were talking about that, I remembered the phrase “back in the day” and mentioned it about 12 months ago. What I meant by “back in the day” was the year 2000.com boom. Maybe with the pace of change, we mean “back in the day” was six months ago. So we have to be careful what it means. And I think there is a time for boards to get ready to invest, throw away and invest, make mistakes, and go again. 

Suppose you had one piece of advice, Sultan, to organizations wanting to integrate these wonderful technologies, the AIs, the ARs, the VRs, the Web3s, the Metaverses, the Conversational AI, and the Blockchain, into their enterprises. What would that one piece of advice be?

Sultan Mahmood:

Yeah, that's a very good question. So, the one piece of advice for me is: don't try to get everything right up front. Don't try to do analysis paralysis. Yeah, it's a given that these technologies can improve. You probably don't know how they can improve yet. So, the only way to do that is to dive straight in, implement some use cases, and learn from that. So that's probably the one piece of advice. We are not in the time now where we can afford to wait and do analysis paralysis and then wait again to see what the results are. Because the competitor won't be waiting, competitors will be moving fast. 

Kieran Gilmurray:

We live in interesting times, Sultan. 

Sultan Mahmood:

It's very exciting. Very, very exciting. 

Kieran Gilmurray:

Someone came up with a phrase, and I wish it were me. They talked about digital "Darwinism," and you've talked about it today as you finished out: If you don't digitize and you don't change. You don't adapt quickly and create an environment where everybody else has to adapt and follow you. You're in danger of turning around like the dinosaurs did and going extinct. We have amazing technologies these days, and I adore all of them. As you mentioned, it may not be throwing everything at the wall simultaneously but picking the technology that reinforces or augments your business strategy or redirects it. And I love your advice, which every company should listen to. If you don't understand the technologies, then go to a company that does because it may look like a cost at the beginning, but when you don't do that, the amount of time, effort, and money that you do waste is an exponential number compared to that original figure.

Sultan, thank you so much for coming to DRUID Talks today. I love the model. I love that there's a plan, place, and time where organizations can invest and move forward. And I love your advice, which is to go fast and create your opportunity as opposed to waiting for an opportunity.

Sultan Mahmood:

You don't want to be the dinosaur because this is a Darwinian moment for organizations.

Kieran Gilmurray:

Fantastic. Thank you, sir.

Sultan Mahmood:

Thank you very much.  

The next episode of DRUID Talks is scheduled for the 31st of January. Subscribe to be notified at https://druidai.com/talks.