[00:00:02] Rajan:  AI is reshaping businesses, and ignoring it could leave your industry disrupted. I'm Rajan, and this is Pivotal Clarity. We talk to those building or using AI, founders and engineers with real world experience. Our aim is just to cut through the hype and see where AI is truly making an impact. If you're a business or following tech trends, these conversations offer clearer insight than most of the press. Let's get into today's episode. Welcome to the Brutal Clarity and AI podcast. Naveel Siddiqui is our guest today, and he is the global head of tech and automation at SAP Digital. Naveel has worked across the industry for over a decade and has seen the tech waves across the cloud, deep learning and AI now. He's trying to weave AI into the fabric of big companies, not as a buzzword, but in ways that solve real problems. I'm curious to hear what he has to say as he is in the trenches with AI and enterprises. Welcome to the podcast, Nabeel.

[00:01:05] Nabeel:  Thank you so much, Rajan, and congratulations for you to set up this platform, which is, uh, very useful for founders, which are very useful for companies who are trying to adopt technology and who are trying to do something new. So congratulations for you for building this platform as well.

[00:01:19] Rajan:  Thanks, Nabeel. How did you get started in tech? How did you move into AI?

[00:01:24] Nabeel:  Very interesting story. Uh, I have over a decade of experience. Did my bachelor's in computer science, and that's where the spark of machine learning started and doing something with data started long back when I was studying computer science. This was not the time, uh, when AI was popular or in the news. So a lot of the work that we did was not publicized like it is being done right now. But that set a very good foundation for all the things that I'm doing right now because the science behind tackling data and building intelligence has not been developed in the last couple of years. It could be very different popular opinion, but this is going on for years years, and people have been building machine learning algorithm. Right now, the amplification has because of the new tech trends that we are seeing in the couple of years.

[00:02:11] Rajan:  What has changed in the enterprise with this new onslaught of AI?

[00:02:15] Nabeel:  A lot has changed. Specifically, I would say, AI has started to come in the process of strategy, operations, And the methodology is that how company used to operate in the past and now has completely changed, and AI has become a central focal point when you talk about even in boardrooms, even in discussing strategies, right, which was not present, like, a few years ago. Right? So that has changed for good because, uh, as a company recently predicted, right, it will be a $900,000,000,000 industry by 2027. So it's a right shift. We are seeing amplification of use cases. We are seeing adoption of use cases, which is all positive indicators for the future of AI.

[00:02:58] Rajan:  Nabeel, you've worked in machine learning, deep learning before, and now it is generative AI 5 years. Before, AI was very different. What is the difference that you foresee?

[00:03:07] Nabeel:  The kind of use cases that we had few years ago were very centric to a particular industry. So when I was working with IoT and, uh, predictive analytics a few years ago, it was centered to, uh, let's say, making, uh, the maintenance of windmills smarter or the maintenance of oil rigs smarter by looking at their data and then giving an analysis. Right now, AI is in everything. Right? It's not specific to a few industries. We don't have any early adopters anymore. Everyone is an AI adopter. It works for a sales colleague. It works for all the roles and functions within the organization as well and applicable to every industry when governments are adopting AI now.

[00:03:49] Rajan:  What is the difference that you see, uh, within an enterprise? Uh, one of the things that happens is whenever a new tech trend comes up, you know, we've we've all heard about the cart and height. So what you're seeing is is different between what is getting hyped up and what is real with an enterprise?

[00:04:04] Nabeel:  Very good question, Rajan. So the hyped up piece is that everyone is thinking that AI will work for all the use cases. So that is the assumption that everyone is going with. That I feel is not correct because AI only works if your internal processes are right and your apply and your application of AI is right to your workflows and processes and use cases. Just because there are reports and analysts are predicting that AI is gonna be the next big thing and it will stay for long. It doesn't mean that it will work for your company or for your organization. And it does also doesn't mean that every use case, AI can be fitted in into everything. Here also, as the same goes, one size doesn't fit all.

[00:04:47] Rajan:  You meet, uh, founders in your

[00:04:49] Nabeel:  day to day work.

[00:04:50] Rajan:  And these days, they are all focusing on building AI products. And, uh, they come and talk to you about, like, maybe partnership or maybe even some advice. What are mistakes that you see that founders are making when thinking about building AI for first products and, uh, wanted to work with a large company?

[00:05:09] Nabeel:  So I do meet a lot of founders in, um, the tech meetups, uh, or conferences. One thing that is good to see is that they're investing a lot when it comes to innovation in AI. But as we have seen historically, every innovation does not lead value or does not bring in the revenue for you. Innovations fail. So in the ecosystem of AI, right, there is, uh, there are different players who are doing different things. For example, the big, uh, cloud service providers, they are taking care of the infrastructure. Then we have a lot of companies who are in the software as this place where, you know, they are creating. They're using AI infrastructure. They're using all the underlying AI foundational models, uh, technology to build applications which could be used by end users. So in this ecosystem, everyone has a specific role. So if a founder is starting a company, they should know where they fall in this ecosystem, which is at times when I talk to them and track in the Venn diagram, they're often found in between. Right? They're not sure whether they want want to invest much on infrastructure or creation of the applications in the SaaS model. That and also choosing the right partner. Because if you are starting a company in AI, you would need infrastructure partners. You would need the big service providers to provide you the models, to provide you the computing power to run your applications. And that has to be, uh, the first step even before, uh, going and evangelizing your idea to other people. Because if you don't have that infrastructure, no matter how good or bad the idea is, but you will not be able to implement it. So strategically knowing the place in the ecosystem. Yeah. That is what is missing from my interaction and also, you know, knowing who's your customer. Starting with the right steps. Starting with focus group study and understanding the needs. Then going the into the direction of validation without just assuming that it will work for everyone.

[00:06:59] Rajan:  So in a fast moving space, like, you know, AI where the ecosystem and the map keeps changing, how does one keep track and how does one keep up to date for themselves?

[00:07:09] Nabeel:  We have a lot of, like, analyst reports which are being generated from, you know, Bain, from McKinsey, Gartner, as you rightly mentioned, the hype curve. It's very important to follow the trends industry through these. Even the big enterprises, they associate with with these analysts so that they can guide them onto the right path. So knowing the ecosystem, and you can only know this by reading these reports, by going to conferences, attending sessions of different companies, doing a good market research to understand who are the players in which sectors in this ecosystem, which are defined very properly and, uh, extensively now after 2 or 3 years. So there there are a lot of, like, ample resources available for this.

[00:07:51] Rajan:  What's the most interesting use case that you have seen inside an enterprise using AI?

[00:07:55] Nabeel:  I have seen a lot of, like, workflow use cases. So I have the pleasure of, uh, leading one such workflow use case in the organization that I belong. And what I'm seeing is that in your internal processes, if your internal processes are set right, be it in your sales organization, if you are putting AI into your internal workflows, there's a huge increase in FTE productivity that could be seen if if done right. So in my experience, empowering the sales executive or account executive who are working in their their life, prospecting customers, writing to customers, they need a lot of, like, support in their job because, uh, they have high targets. They have high targets to meet every quarter. And manually at times it's not possible because data in a big organization is scrambled through different sources. Right? You have excels. You have applications. So if they have a AI platform where information is on their fingertip when they are doing prospecting, when they're doing their interaction with the customer or doing research on a particular account, that is very fundamental to the growth of any com any company because that's where pipeline is built. That's where revenue is generated. And a successful AI integration there could bring in 1,000,000,000 of dollar of impact.

[00:09:12] Rajan:  I spoke about, like, you know, an internal workflow. One of the things that I've seen is as an AI, you pick the right use case, there is crazy growth, and you don't pick the right use case, then, you know, there is almost no growth. So between internal and external use case, and maybe you can talk about, uh, some broad strokes of categories in, say, in the HR or finance, what do you think is the sequence of how these are getting adopted? I mean, today, is sales getting adopted first? Is marketing getting adopted first? And what do you think is gonna happen in the next few years in terms of, first, this will this use case will be adopted? If you wanna take a view of what would it look like?

[00:09:49] Nabeel:  In my experience right now, the relevance to sales and relevance to people who are interfacing and prospecting is good because they need support in writing emails. They need support in writing the call script for the conversation with their customers, objection handling, and so on. So the current models that we have and their, uh, capability aligns well with such use cases because, hey, we all have been using Charge GBD, for example, to write emails. Right? Because it it it works well. It knows how to write email. But if I expect the Charge GBD to support HR in their HR workflows right now, Maybe it would be a stretch because the intelligence and the inference part is still yet to come or it's coming in batches right now. So decision making is something where I think a lot of investments would be needed on the infrastructure, on the models. But if the right set of instructions are given, hey. Then I want to write this email to the prospect. The prospect is here tied to the internal data. That's where the wind is right now with the infrastructure that is available. The future, uh, would be that once the technology improves, as I talked about the, uh, ecosystem. So everyone in this ecosystem is dependent on one another, and there's a bidirectional contribution. Right? The more the use cases, the more would be the need for the investment on infrastructure. Once this infrastructure is available, AI could be amplified to other rules, uh, to decision making support as well. To management support level as well, uh, where scenarios could be played or simulated on an AI, and then AI could tell you where in which direction you should take decision. So I feel there's a huge, huge, huge opportunity, uh, in this ecosystem if things are implemented. If this value chain, as I call it, is, uh, implemented, and it keeps on growing organically as it is predicted.

[00:11:35] Rajan:  You mentioned about, like, Tag GPT, uh, models. Is there a preference of models within enterprise, hosted versus owned versus open source? Like, are there, like, you know, thought process that are emerging that within the enterprise, do you see any trends on what type of models are getting more adopted and what type are getting listener?

[00:11:55] Nabeel:  I think everyone right now is doing a multimodal approach where every model is available in all the organization whether it is LAMA, whether it is the OpenAI models or model and so on. But now the organizations are realizing that, hey, there could be some use cases where a particular model could fit in well. For example, if you're creating video scripts for to generate product, uh, launch videos. Maybe Llama could be better in that, not the OpenAI's GPT. But if you're writing emails, maybe a particular model can outshine other models. This is where now the analysis and the research is going on within organization that which model could do which task in their ecosystem. Then I'm also seeing a lot of imaging trends in in the industry where companies are also training their own models. They're using RAG. They're using, uh, the, uh, vector DB approach to, you know, have their knowledge, uh, portals be trained on some models so that the FAQs and q and a's that customer ask on chatbots, they could get good answers because now the technology is there. RAG is there to vectorize your database to have notes created for all the data points with the right keywords. So that if you ask something where we have knowledge, the right answers could be given with references.

[00:13:10] Rajan:  None of everything is as hunky dory. While there is hype and there is excitement, I'm sure that there are failures also in terms of projects. What are some failures that you have seen recently, and what are a few things that can be learned from those?

[00:13:22] Nabeel:  In general, as I I would like to tie it back to what I said, uh, at the beginning of the podcast. If you're not having a right process in place, which is the foundation for the implementation, no matter how much AI you try to fit in in that process, it will not improve your productivity. Right? So if you have a bad process, then AI cannot help you. I'm seeing, uh, you know, some use cases are being tried on the current models, which are ahead of time. Because the expectation because of the hype, the expectation all also goes up of the user. So people are trying out, as I said, decision making. Uh, so they are asking AI to do decision making on behalf of them in several organizations. So that is not what where we are right now in terms of maturity. And projects like this, they're not I'm I won't say fading, but there's a good learning in those because, uh, these will become future pipeline of use cases for AI. But this is where we are seeing not as good of a result as, uh, we have seen in other use cases.

[00:14:19] Rajan:  One of the things about enterprises is, uh, how purchases happen. Mid market companies have different behaviors in purchases. Enterprises have different behaviors in terms of how they buy things. What are the things that are changing because of AI on how enterprises are purchasing? Just to add a little more color to the question, um, earlier, like, you know, the headcount budget and then the software budget used to be different. But now due to AI, there are talks about that getting overlapped or, like, you know, one is seeping into each other. What are the things that you were seeing in terms of how enterprise behavior of buying has changed?

[00:14:54] Nabeel:  It has changed drastically because you have also been part of a big organization, Rajan, before, and they bet on value when it comes to purchase. So now because the value is being predicted in 1,000,000,000 of dollars, uh, $900,000,000,000 by Bain, we are seeing a lot of purchases happening in procuring new models, in procuring or, uh, you know, acquiring start ups, which are already making headways in, uh, AI. So this was not present before. Acquiring the best model or acquiring the best AI team, which is present out there, was not a priority of purchasing before. But overall, industry wide now, everyone wants the best model to be available for their organization. They want the best models to be embedded in their products because then that also trickles down to the revenue, and customers are also asking for it. So they have to do this. So because of the value that peep everyone is seeing across use cases, across organization, across industries, there's a good impetus to invest more in AI, which is good for the future of AI as well. And this also works well with the ecosystem that I spoke about. So now the money is going back to the big cloud service providers to invest more in the r and d and and bring out new exciting features.

[00:16:10] Rajan:  Nabeel, I'm gonna ask a couple of quick question. Uh, tell me the first thing that comes to your mind. What is the biggest misconception about AI inside enterprise?

[00:16:18] Nabeel:  That it works for all.

[00:16:19] Rajan:  There is this debate about, uh, big tech versus start up, which will win in AI. I mean, every platform shift, you talk about cloud, then it is mobile. So similarly, in the AI era, who do you think will win big tech or run start?

[00:16:32] Nabeel:  Again, depends on where you are in the ecosystem. If you want to compete on the r and d, it would be a very tough battle to do with the big cloud service providers. Right? Because they have a huge infrastructure to support con continuous innovation in that. But if you are trying to win with big big organization on, uh, on softwares and on applications, that's where I think emerging technologies could give a very good competition to big tech. So depends on where you are in the ecosystem.

[00:17:00] Rajan:  Favorite AI tool?

[00:17:01] Nabeel:  I'm biased to say my own tool there, but because of enterprise restrictions, I cannot name it.

[00:17:06] Rajan:  Future of AI in enterprise.

[00:17:08] Nabeel:  Huge potential given what's coming in terms of infrastructure support. Wide adoption across rules, across pillars within the organization, be it HR, be it strategy and operations, be it customer success or sales organization or marketing. Huge potential to have the complete 100% adoption in all the pillars of an organization.

[00:17:31] Rajan:  Is there a non obvious, uh, use case of AI within enterprise that most people are not paying attention?

[00:17:38] Nabeel:  It's been there for quite some time. So a lot of the use cases are now, you know, out in the open. But I think, for example, microsite generation. Right? If you're creating a website, if you need right content to be placed in different sections of the website, we create these microsites for customers, all the organization do. I think there, it's a non obvious, uh, because people don't think that how AI can help in generating content for of site that has been we'll be able to promote a particular product to the customer. So that's where I think it's a bit non obvious. There are obvious winners here as well, emails, call task, and all video scripting. But, yeah, this is microsite would be one of the non obvious things.

[00:18:18] Rajan:  Coming towards the end of our chat here, Naveel, imagine 2034, AI and enterprise is at the widest adoption that we have imagined, uh, today. How does a day in an enterprise look like?

[00:18:31] Nabeel:  Okay. We are in 2034, and let's assume that AI is everywhere in the organization. Right? So I think a lot of things would change in terms of how we execute tasks. Right? So if you would see a salesperson doing his job, you would have a dashboard exactly where, uh, he would know which customers are ready for upsell, which customers are ready for new product talks. And, uh, people will stop shooting in the dark. Right? Everyone would know where the value of their effort should go in terms of realizing a great benefit. So that's what would, uh, the AI would change. There will be no hidden trial because we'll have a lot of information. We'll have a lot of simulations run by AI of different scenarios. And I think people will start hitting bull's eye every time they're trying to shoot a target.

[00:19:19] Rajan:  So around compliance, especially in large companies, if you don't have SOC 2, you cannot get involved, and maybe because of AI, they might give you compliance that will come up. What are the things that are important as of today? Is it the regulation or is it, like, you know, something which is more internal compliance and governance?

[00:19:35] Nabeel:  I think both because compliance are there for a reason. Because if the data that you are collecting as part of your fundamental of building the AI has to be a clean, compliant data. Otherwise, the whole process is not done right. So I think compliance is very critical. But using technology so that you can, uh, anonymize data, so that you can build confidence that, hey. I'm not gonna use your data blindly. I'm gonna ensure that there is a sanity towards, uh, the usage and how I'm procuring data and everything. So that cleanup is required. And for that, I think compliance is necessary, and that's why business AI but responsible business AI is where organizations are going in, which is very good to see.

[00:20:20] Rajan:  Abhil, thank you so much. This is such a wonderful conversation. Thank you for joining the the Vertu Clarity podcast.

[00:20:26] Nabeel:  Thank you so much, Rajan. And, again, you're doing a great job doing such a work because this is giving a lot of information to your listeners, so thank you so much for doing this.

[00:20:35] Rajan:  That said for this episode of pivotal clarity. This is an Opeka podcast. Opeka is an accelerator for global Indian founders building AI software companies. We're exploring the fast changing world of AI together with our listeners. If you like this podcast, you can find more on our website and other popular podcast apps. Subscribe if you want to keep up.

‍