[Music] [Ben Stobart] Welcome, everybody. I'm Ben Stobart, and I'm joined by Ravi Pal. He's Global CTO of Ogilvy One. Today, we're cutting through the noise and to talk about AI and what it really means for enterprises.
So we want to make sure that we're covering the perspective of brand, business, and customer relationships. Ravi, as we know AI is at an inflection point.
Years of digitization, automation, we're seeing a platform decay, someone describe it, where you've got prioritized short-term gains over long-term integrity, whether that be brand or where the business is going as itself. But as you start to look at where AI is positioned, what's your view of enhancing consumer trust rather than eroding it? And then how do you actually do that? What are your thoughts on that? [Ravi Pal] Hi, Ben. - Hi. - I think it's a very timely conversation. As you know that we are also going through our own AI journey as an enterprise. Every organization, every brand will have their own journey, unique journey. But I think the core principles apply so you're right. Sometimes we can be swayed by the euphoria of new technology. AI is happens to be this profound new change that we're all witnessing. But I think if we go back to simple basic principles of how we made decisions for the best of our business, we could still deal with the concepts like platform decay, etcetera. So if you think about it, customer first and technology second. So in this case, AI second, so you can still make the right decision. Ask yourself the question. If I use AI, am I using it to better the customer experience, how customer interacts with my brand? You can look at your employee augmentation. So think about how you're going to extend the human agency for your employees so that they can partner with you in achieving your business objectives. And then if they have to deliver value, then you need to define value. So what that value is. So for any automation technology, you would imagine productivity is a byproduct, right? So that's there. But how do you define additional value? How do you create some ROI gains for your business or your overall objectives if you're a marketing organization? So I think those three core principles, that has always worked, and I think that should work with this change as well, but it has to be underpinned by the transparent, ethical, secure use of AI. That'll give you this holistic set of principles to allow you to navigate through this change and ambiguity. Yeah. Absolutely. And I know you and I have discussed this a lot, that while you can talk about short-term and long-term in this ideal position of where AI is, this does seem to be two tiers of businesses, one where adoption is happening quite quickly and successfully, but where others are falling away. So I wonder while we talk about this conversation mattering now, which obviously it does, every conversation has got this in. What are the key steps an enterprise needs to take to future-proof the business? Maybe you could touch on that first point that we were discussing earlier in the week, which is understanding whether is AI the problem or is it something else? I think-- I don't think AI is the problem. So let me, frame this in broad buckets. Yeah. So think about it this way, right? So any transformative shift in technology leads to, for us to ask these questions that has happened in past when cloud came through or even when internet came through, right? So the three frames that we should look at are future-proofing your organization in what sense? Look at the roles, look at the processes, and look at the ways of working, right? So think about from the employee first perspective. So that's one frame. The second frame would be re-look at your data strategy because most of the companies by now have a data strategy. They had identified this in last decade or so that data is going to be the new oil. So they do have, and every company has a data strategy, but they need to relook at it with the usage of AI or the reintroduction of a new technology into their landscape, which relies on good data. And then the third frame will have to be, so how do you use the core aspect of your business, which is being customer centric, right? So how do you create growth by being customer centric or designing relationships with our customers, and between brand and customer, which are unique for your customers? So I think that's how I'll frame these principles that should help us address the question you had. Yeah. So do you want to take us through those three pillars? Because certainly the one of the ends which I took away was about relationships, but what about the first pillar? Because I think that when we talk about adoption and who is adopting early, why does a gap exist when some are adopting very well and some aren't? Do you want to touch on that first piece? Yeah. I think there are general scenarios and circumstances. So keeping that aside so that we don't sound too generic and try and keep it a little bit more specific.
If you think about it, there are two spectrums. We can't touch on everything. So let's take two spectrums, right? So there could be a customer who's waiting on the fence. They want proof of value to emerge and then dive in and see how they are going to take advantage of this. And there could be a customer who's looking at it from right the trend, get ahead of the competition, and create enough differentiation so that they can win the race, right? So in both cases, you would still have to go through the core changes or the principles if you look at from future of organization or future-proofing your organization standpoint. Why should the roles change? Why should the process change? So let's say, in one case, you buy a technology and you want to bolt it on top of your existing process. How's that going to affect or create value for you? If you had a bad process, you're just automating a bad process. None of us want to do that. I mean, you and I in our day-to-day go through, we question our processes at times, and we want to relook at how the work gets done and not necessarily there's a process, let's automate it. So you want to unbundle your current process, think about the core steps of how work gets done, and then look at where the infusion of this technology is going to create some efficiency, some productivity gain that will lead you to look at, so what should the new roles or roles that needs to change work as? So here is an example, right? We were working in this creative space. So take creative as a discipline. So today, creative comes up with an idea, but then they also have to run with the process of creating several variations and scale them out. If there are geographies involved, then you have to create several translations of the creative output. But if you change that process in the right way, the creative could now focus on perfecting that idea, create that unique on brand creative that your customer is going to like. And then let the technology take care of creating probably variations, creating translations, creating or solving the scale problem for you. So that's how roles and processes would change. But I will add one additional thing, Ben, that while the workflow processes need to change, ways of working also needs to change. So with the same example, when you unbundle, something and you look at where you need to infuse intelligence into your process, you also want to figure out then how do you validate that change? Because this is new, so we shouldn't lose the sight of, this is a new technology, and the proof points are limited. So for you to quickly validate that, prove that there is ROI at the end of this, and then scale it. So having that delivery process or a rigor that works for you as an organization is equally important. So in summary, I think the vision to execution, these are few areas one should look at to be successful.
In summary, we've spoken about this a lot, the change management of opening hearts and minds of people to absorb. That seems to be one of the parts as well, the ability for people to embrace the benefits of AI. And AI is somewhat dependent on that human aspect coming in at the beginning, which I think is one of the successes of change management and one of the successes of policies being changed and approaches being changed. It can only be changed if people are open to them. So I think that's a really fascinating part of the journey.
Moving on to, again, something very technical, but also very, I suppose, relevant to the people element of AI, which I think is constantly the discussion. So data silos. So you make this phrase, data silos aren't technical, but they're political. And I think that's quite an interesting view whereby you can have this perfectly tied up box of this is our data approach, and this is what we should do for good governance. And this is our approach for how we're going to use it. Here's the vision. But taking data and moving it into an end valuable product or entity is very, very hard. Do you want to expand on that, the second pillar of data and how silos are, while technical rooted in not at the time in company politics? Like I said, I think we need to relook at as we are doing, all organizations, all enterprises need to relook at their data strategies. So data problem is not new. I remember in 1999, first time I worked on data mining problem. And we've seen several iterations of innovations in data space from big data to data lakes. And even in data engineering space, we've seen innovations.
To me, I think most of the organizations have reached a point where it's harder for them to now find new ways to break those silos. And some of those silos exist for sociopolitical reason, you and I have discussed. Some company acquired another company, so brought in new set of tools, governances, definitions. You have different departments. Their incentives are not aligned, so they continue to work in their own silos. There is operational data. There is strategic data. So there is a genuine reason why some of these islands and silos exist. What we need to do, when we look at AI is relook at these data governance strategies with a different eye? Because your data preparation changes, your AI is as good as your data. So you need to prepare data. Your data needs to be clean. Otherwise, you'll be spending wasting a lot of time and resources and training an AI with a bad output. You need to make sure that the data is labeled correctly for the models to be trained or in some cases, fine-tuned to work for your cause or your input. And then finally the new spectrums of compliances that come with AI. So that's one aspect of it or one pillar of it. But then there are technical challenges. So you don't want to now reopen your entire data strategy. You probably want a more holistic integration strategy. So take as an example-- I'll take couple of examples, right, so that it becomes real. You have, let's say, any customer today has a data lake. But for AI, when you talk to a knowledge repository and you want to discover or retrieve information, we call it search use case. Then you have lot of these unstructured knowledge documents available coupled with some structured or semi-structured data stores. To build that, you need new plumbing infrastructure on data. So where does that sit? So you're going to augment your current entire data hub strategy or data lake strategy with the new tools and technologies for you to enable this new application. On the other hand, you can have a scenario where, if I take the tool that we are building internally for our teams, you may be looking to find something which just becomes a function call, and it could be just an API call to an existing system. So you don't need to break anything. You just make sure that your application has an API, which is well defined so that your AI can just do a function calling. But if you had a complex query and you wanted for your agentic architecture to then dismantle that query into reasoning, planning, and then identify how it is going to go talk to different data stores, come back, resummarize that, and then bring it back to you. That becomes a complex use case. And for that to work, you should have trained your AI to be able to reason, to be aware that how is it going to go and talk to different data sources. So as you go on that spectrum, your use cases become complex. But that's where the value is, and you will have to unlock that value eventually. So that therefore, relooking or inward looking for how you're going to enhance and improve your data. I call it a new integration strategy where you probably have to bring several data sources together instead of recreating another platform. And then bolt on top of it, a federated strategy where data products are still owned by, your social political structure allows for marketing to own a certain database, sales to own a certain database so they can act like a data platform owners, but then make sure that is accessible and available for people who are creating these new valuable use cases for you.
Really interesting. I mean, if you are to look back at probably some of the biggest core challenges of our clients and even ourselves as we've gone through this journey is that it seems to be the same old challenges that we've had, which has been silos, one, unlinked strategy or a strategy that doesn't link everybody together to have a common goal, and sometimes compliance potentially getting in the way of entities coming together through no fault of their own.
I just find it interesting that the same challenges that come up traditionally are also coming up again now, which I don't think we can get away from, it would fall for those that are going to be successful in this space. So I think the shared accountability is something that really is going to drive this. So it's funny, isn't it, how the same things keep coming up? Yep. Just wanted to move on to, as probably a more favorite subject of mine, which is about the end result when we're talking about, this is great. So how are we actually having return on investments? And you and I have talked about the tangible strength of let's not just automate. Let's not just look at this as an efficiency and forget about it approach. AI needs to strengthen customer relationships. That's actually the benefit, not just automating and taking out effectively layers.
What's your view on that? And then more importantly, how departments are able to take this back and show genuine growth, which shows genuine return on investment, which has sometimes been a little bit of a hindrance to the MarTech space in the last three to five years. Do you want to talk a little bit more about personalizing experiences, but at scale and how brands might struggle to use, sometimes users effectively if they're looking at it as just an automation brief? Yeah. No, I think that's the crux at the end of the day whether you retrain your employees, create new roles, or redefine your processes. End result is we want to make sure that we are able to create these strong bonds with our customers.
I think the personalization, or creating these unique bonds, which goes beyond personalization as well because it's a different gradient where you are in the lifecycle. So you may be looking at it. Let's say if you're in an acquisition frame, you may want to look at what dynamic content I can create using AI? But can AI also help me predict the scoring of that dynamic content for a given segment for a given channel so that I know that I'm not creating inventory of waste, and I'm creating something that actually is going to work? You can look at it from a commerce lens, and you can say, "I've got 100 SKUs." And I've got, let's say, 100 different personas that I need to probably serve these product descriptions or SKU related information too. And if you go through channels, etcetera, so these become thousands of variations, right? So how can AI help me create unique product description as Ravi wants to see or Ben wants to see, even the imagery or creative around that product, the tone and tonality of that product description? So can I help me create that scale or scale the personalization, right? And then we talked about the translations, etcetera, if you add the geography or market regions to it, right? The third is look at social space, right? So how can you be ahead? And let's say, there's a solution that we're working on as you're aware Influencer Shield, right? So every brand has influencers who are great, right? They bring in customers and helps with the acquisition. But what if any comment creates negative publicity? So can AI help me identify that in time, create an alert, and let the brand or brand manager know that this could lead to a negative publicity and help me nip that in the bud? So all of these different scenarios where you can use AI to address the personalization problem. If you switch to from more B2C cases to B2B cases, and B2B, we've looked at personalization very transactionally, right? So if today, you and I get bombarded with lot of our vendors trying to reach out to us for new technology tools, and you will see a very generic approach. And both of us will be given probably the same message irrespective of our roles or irrespective of our interests, right? But within B2B also identifying that buying group, identifying them as individuals, their interests, and then crafting personalized messages that work for them depending on their roles and their interest and their intent, as well as on the channels that they are on and would like to receive that message on. If you're a football enthusiast, is it being published on one of those sites where you are, right? So I think we can look at both spectrums and then use AI to genuinely create a differentiation for the brand and businesses, which really want to create these new avenues of forging unique one is to one bonds. Yeah. Look, you and I can talk for hours on this subject, and I know that we're being right at the coalface on having to deal with this. Last thought. When we talk about strategy, we talk about design. Have you got literally one minute for us to talk a little bit about how you would approach this? We look at this with relationship design within Ogilvy and understanding all of the facets you've talked about. But have you got a one minute closing, Ravi, for people to focus on when we talk about design and bringing these goals together? Yeah. So relationship design is our unique proprietary methodology. That's our philosophy, way of working. But truly, what it is that it allows us to create a unique proposition for each business. So what does the brand wants to speak to their customer? What is their business ecosystem? What does their customer want and why they want to buy from this brand? What data they have? So how do we create value using that data or technology investments they have already made? And then come up with this relationship platform that creates unique differentiated outcome, communication, creative, platform, journey so that we can achieve the objectives that we can. But it all starts with what I call an assessment. So as a business, you may want to start with saying, are you ready? Have you looked at all aspects of your AI strategy? But as we talk about people, process, ways of working, have you looked at it from your data infrastructure, your technology infrastructure, even architecture blueprints that would you be able to introduce a new technology into an ecosystem where your IT teams may not be ready? They may they may have to develop new control centers being able to create trust with the business that we can identify the metrics that allows us to learn about how this model or how this AI is working in the production so that you don't have the trust deficit or erosion of what we call the trust with your customers by creating something negative, or putting something which is not a quality product out there, right? So I think those are the two cracks, but relationship design at the center of it is our way of doing this for our customers.
That's great. Thank you very much. I know we're going to be talking more about this at Summit, as we come up to that in the next few weeks. Yeah. Thank you. And like I always say this is an important topic for us, as well as for all of our clients. So this space is evolving at a very rapid pace. And I'm sure we will continue to have more discussions around this. Thank you. Thanks once again. Thanks, Ravi. And thanks to everyone who joined us. We look forward to seeing everyone at Summit. Please come up and say hi to us. In summary, making our AI work is going to be work, but not doing anything is going to be critical to your business. So as we say, let's go beyond the hype, make AI work for us, and not the other way around.
[Music]
