Using AI to Revolutionize CX

[music]

0:00:03.7 David Chao: Welcome to Data Radicals. I'm your host, David Chao. In this episode, we are thrilled to be joined by Michael Olaye, EVP and Managing Director of Hero Digital. Michael shares his incredible journey from aspiring pilot to tech industry trailblazer. He reveals how he's using data and AI to transform customer experiences with real examples and valuable insights. This episode is a must listen for anyone passionate about the intersection of AI innovation and customer experience. Stay tuned.

0:00:33.5 Producer: This podcast is brought to you by Alation, a platform that delivers trusted data. AI creators know you can't have trusted AI without trusted data. Today, our customers use Alation to build game-changing AI solutions that streamline productivity and improve the customer experience. Learn more about Alation at A-L-A-T-I-O-N.com.

0:00:57.4 David Chao: Today on Data Radicals, I'm joined by Michael Olaye, EVP and Managing Director of Hero Digital, where he drives client strategy and business objectives. Previously, Michael served as the SVP and Managing Director for Strategy for Innovation at RGA and held various leadership roles at national and global organizations in roles such as Chief Growth Officer, Chief Technology Officer, and Chief Executive Officer, working with brands like BMW Group, Credit Suisse, JP Morgan & Chase, and Unilever. Michael, welcome to the show.

0:01:25.7 Michael Olaye: Thank you for having me, David.

0:01:27.3 David Chao: So I just gave a little bit of a brief intro, but I'd love to hear more about your role right now as a Managing Director at Hero Digital.

0:01:35.3 Michael Olaye: Sure. I've been at the organization now, I'll be coming up to four months. I lead the Western Region of Hero. It's a customer experience specialist company. We don't like to say we're an agency. We also don't like to say we're like a consultancy. We're like a hybrid. And we basically, service big brands and small brands across customer experience, which is around everything from data to products, to platform to services. And I run the Western region. So I like to think everything from Wyoming all the way to Japan. Everything there is in my region and I basically work across the entire business literally. So I'm the portfolio owner and the executive sponsor across all our accounts. I help with growth, I work on growth a lot, retention of clients, retention of talent, hiring, strategy, vision of that region. And I'm also on the executive management team as well. As we scale as an organization, I partner with my colleagues to make sure we're doing that properly.

0:02:25.2 David Chao: That's amazing. I know you started your career as a software developer. How did you make your way from front-end web development all the way to the role today at a creative agency?

0:02:35.4 Michael Olaye: I think I just fell into these roles. I actually studied to be a pilot at a university. That was my dream. And then I learned HTML from the HTML for Dummy's book, believe it or not, as a side reading thing. I actually got my first role in the industry by knocking on door to door. I actually built a map of London underground and I found all the agencies and all across the map and I built a circle and I literally left my house at like 7 in the morning and just randomly knocked on agency doors. Some agencies said, go away, some agency let me get as far as reception. A few took me to their studios. Yeah, and I just tried until I got my first job. And my first boss was a creative director and art director in a tech company that had a design studio. So I actually started off in a design studio. And I was in a lot of front end stuff obviously, and then ended up doing a lot of backend things as well. So I ended up building a skill across like using programs like Illustrator, Freehand, and Fireworks. This is a pre-Photoshop, and then also writing coding director, but also like Action Script, being able to use keyframe animation, timeline animation. So, I ended up learning both design and coding at the same time, which was useful.

How do you use data for customer experience (CX)?

0:03:37.3 David Chao: Amazing. That's such a tale of perseverance. I'm imagining you going on the tube knocking on these agencies and getting turned away. And so fast forward 20, 25 years later, you're now at Hero Digital at the center of customer experience. And one of the things I know we wanted to discuss today is the role of data in understanding and shaping customer experience. Can you tell us a little bit about how you use data in your day-to-day role?

0:04:01.8 Michael Olaye: Yeah. I mean, we use it both internally to make like business decisions, everything from how our employees operate, where we invest, all the way to like building things for clients. So whether that's insight gathering, whether it's actually building a product and using data to actually make those decisions. And then also when we launch things as well, we're constantly getting feedback or you're constantly like tracking the data against what you've done. We also have a kind of media arm as well. So we do a lot of things in that space as well. So I would say for us, data is at the heart of like the organization, both like inwards as well as working with clients and brands as well.

0:04:35.7 David Chao: And when you think about customer experience, it can seem very intangible, very fluffy. What do you think is appropriate to measure and not measure with respect to customer experience?

0:04:46.6 Michael Olaye: Yeah I actually don't think it's fluffy. I actually think it's very straightforward, right? Like depending on the vertical you sit in, the experience and the maturity of the experience will vary. So, if we take health for instance, when you wanna measure in health, there's a lot of regulations around that and there's a lot of kinda control of what the data you get and what you can do with it. But actually when you build a product in the health space or when you build a tool or when you build an application, you can track that application, you can get feedback from customers, you can see how that application or product is doing. And you can make, I wouldn't say live adjustments, but you can make like almost close to real time adjustments. So that data is an ammunition for product improvement, business improvement, service improvement, customer service.

0:05:27.7 Michael Olaye: So I actually think the whole CX thing is an offline and online world of how we humans interact with organizations. That is the experience, right? And also internally, how your employees interact with your organization is also CX. So I would say, B2B, B2C internal operations, they're all like CX for me. And a lot of the brands that come to us, they vary. Some are quite young and they're infant and they need us to help them build strategies around what is good CX, what does that look like? Other organizations are very digitized or transformational already. They have a lot of tools, but they might not be getting the return of investment. So they're looking for customer experience from the point of, I have all this data, what do I do with it? So, I think it's quite straightforward. It's just understanding where in the maturity that category is and where in the maturity in that category the brand is sitting in.

0:06:14.4 David Chao: When you think about measuring customer experience, you talked about customer sentiment. Is there an example you can share in terms of how you've thought about measuring customer sentiment?

0:06:22.5 Michael Olaye: There's some good ones I can think of. I mean, I think we talked about the one with Watson, but if I remember, which is basically it's a project I was lucky enough to work on where we had the whole Universal album catalog available to us. We also had Twitter at the time, and we were able to basically scrub a person's profile. And depending on how you behaved at certain times of the day, what you liked, what you tweeted, what you shared, we could build like a personal like analysis and then we could recommend either universal films or music to match at specific time of the day. And actually it was a really good way of using data to almost, I wouldn't say, measure the person, but actually create an application that provided a kind of pleasure to someone or actually made a change to how their daily life, right?

0:07:08.3 Michael Olaye: And I think that's a good way of using data. In terms of measuring data that's already out there, or products that's already out there, we do a lot of reports. We have a salesforce partner, we work with the likes of Adobe. We have a lot of the tools that we use to kind of track data, things from workflows to supply chain, to product development, to media. And we are tracking that across a variety of different clients at the moment.

0:07:31.7 David Chao: And when you think about measuring these insights, is there a limit to what can be measured, do you think?

0:07:37.4 Michael Olaye: I think so. The way we work at the moment is rubbish data in, rubbish data out, right? So usually there's a lot of human error in data. There's a lot of what I would call, especially in unstructured data, there's a lot of useless data. So, it's around cleaning that data, well, first of all, creating a strategy around where you're using the data for and then being able to kind of clean it to the point where it actually is good enough for you to use it as an insight. So I would say most of the clients we work with right now, they have a data team internally, sometimes not too big and they have quite a lot of structured data that we work with. And then I'll say more on the product side, we have more unstructured data. So it could be a market leader say in a banking or financial sector and they're moving into kind of FinTech and a lot of the data that they might be acquiring to make business decisions might not be so clean and so structured.

0:08:26.6 David Chao: Makes sense. I think ultimately with using this data, as you said earlier, to inform decision making, and I know when we talk to our customers, oftentimes they're using these customer experience insights to personalize and shape the offers that they're putting in front of their customers. Can you share maybe a few examples of how the insights that have been gathered from this data are shaping the customer experience journeys for your clients?

0:08:48.8 Michael Olaye: Probably gonna say what client it is, but we have a pretty big streaming cable client. We're helping them currently shift their business from a completely kind of cable-like organization to actually being more of a hybrid, where streaming digital plays a huge part of the organization. They have a ton of data, they understand the audience, they understand the packages those audience buy, but we are also working with new forms of information, new forms of data or what the future demographic might look like. And what does a perfect next best customer look like? So there's a lot of modeling, data modeling going on and we are working to help them create kind of a new business model where they can compete with the likes of the more famous streaming organizations. And I would say a lot of that is around like the data structure itself, but also on a strategic side, how we work with that data to create a business plan and a roadmap that makes sense for the business.

0:09:41.9 Michael Olaye: That could be leading into product development. It could be leading into R&D, it could lead into partnerships for them, but also sometimes it could lead into like an internal organization change. So we know data can provide those insights that usually you wouldn't know if you're inside an organization itself. So, we as an agency coming through and saying, Hey, we can see a pattern happening here or in your industry, this is what the data's saying. And strategically do you wanna play in that field or do you wanna kind of challenge it in terms of how you set yourself up as a business?

How data insights can surprise you: The cougar story

0:10:11.4 Michael Olaye: Is there an example you can share where maybe a company is going in a certain direction and then when they look to the data, they looked at these insights that prompted them to maybe reconsider and go into a different direction?

0:10:21.8 Michael Olaye: Yeah. I did something for Unilever a couple of years ago, which was quite cool. There was a lot of informed data that was, they were using to almost create a campaign. They never went live unfortunately, to try and create a campaign around what does a cougar mom or wife or teacher look like. And the perception on the internet was very different to actually what most women over 55, who are classed as cougars, looked like. And we actually did a really cool analysis with data where we scrubbed Google. I think we've got like almost five gigawatt of like stuff off Google. And we also took images from different universities in the UK of different students, different ages of women. And then we trained a model. We worked with Clarify, it's an AI company in New York where we trained the model to understand and to project what a cougar would look like in today's world.

0:11:06.5 Michael Olaye: And actually what it came back with was actually quite realistic stuff. Wearing dreadlocks, wearing different styles of dressing, looks, some modern, some not. And that actually informed quite a lot of what Unilever was planning on doing across their marketing campaigns for the next kind of couple of years. They never launched that project, but it informed a lot of things that they were looking at and helped them make business decisions on how they should do some of the content marketing on social, how they should do content creation. So yeah, I think that's a good one way, if we didn't do that, we probably would've just sat with the bias and the expectations of what the market thought a cougar was.

0:11:39.4 David Chao: That's a really interesting example.

AI in marketing today: The role of data

Michael, you just mentioned clarify AI and using data to train AI models. And as you said to me earlier, in today's market, you can't have a data conversation without talking about AI and vice versa. So let's spend a little bit of time talking about AI. Is AI something that you are starting to see in your day-to-day work?

0:12:00.9 Michael Olaye: Yeah, for sure. I think the first time I worked with any AI product, 2011, 2010 on the Watson, the sentiment engine I mentioned earlier, that was all the way back in 2010, 2011. And then the Unilever stuff was like 2016. I've worked with data of BMW, there was a lot of more machine learning things. And then I think 2017, obviously when the Transformers came out and ChatGPT came out, AI became this huge thing. And I don't think there's any tool that you can work with today that doesn't have AI. And I got some stats here that I like to call out. So, like over 50% of all US companies with more than 5,000 employees currently have AI, like they use an AI. And if you take that to 10,000 employees, it actually rises to 60%. So, that's any company that has more than 10,000 employees has 60% of the employees using AI. And actually the smaller businesses they was not using AI, so the small businesses, but the larger the enterprise, the more likely they're using it. And when you get to large enterprises, there's 42% of all the employees that are using AI. And we know pretty much every organization has an AI strategy or are thinking about AI strategy. And I think it's the technology that's pretty much fully infiltrated most of businesses today.

0:13:10.9 Michael Olaye: Could be anything from simple use of ChatGPT to Journey to building large language models, to building small language models, to building agents. It is just a whole variety of different things that everyone's doing. And at the heart of all of that is the data. If you have the funding and you have the brand equity and you have this data, you can do some pretty amazing stuff with this technology. It could be everything from customer service to new product innovation, to actually just like changing your entire business model or creating a spool of services that focus very much on generative AI or AI as a whole. So it's pretty exciting times, I think.

AI for productivity versus innovation

0:13:49.2 David Chao: And there's AI in terms of automating or accelerating the work that we do day to day that you touched on, automating customer service or accelerating image creation, for example. There's also AI to dramatically change and reimagine and transform these experiences. Are you seeing AI being used more in the first or the second camp?

0:14:10.0 Michael Olaye: I think AI is split into two right now, specifically because of IP privacy and bias. So, I think there's a whole thing happening over here, which is many companies can do a lot of stuff with AI right now, but won't launch it out because it's just, the data is trained on, there's a lot of copyright infringement in there. Nobody wants to get sued for that. Companies like Adobe and Shutterstock are providing these huge enterprise AI tools that are amazing and protect you a little bit. They also have a limit on the stuff you can do before you get protected. There's like a numerical number in there of what you have to pay. So for many companies, the risk is too high. And there's also a lot of what I would call lack of understanding and education on what AI really means for an organization.

0:14:48.2 Michael Olaye: So I always say any organization that wants to even touch this technology for a like public use needs to have a business like AI ready model. There's a few companies where we're actually working with around like governance, privacy, IP, data source and data storage, data breaching, human error. And there's a whole bunch of stuff. We already know, even with our AI, we've all heard horror stories of companies launching campaigns and things out there without like checking it properly and getting sued. So now to allow a machine to do that, it's too risky. So, that's the first half, but the half that I think is super like exciting right now, which would help the second half later on, is building these internal work process tools, bringing efficiency, speeding up the creative thinking, speeding up strategy. And in our industry in the marketing technology product industry, it's amazing because you can get from like initial concept to like ideation like super quickly.

0:15:40.1 Michael Olaye: You can do stuff that would take weeks before in days. You can collaborate with people who have no technical knowledge on technical things. We have tools now that you can code by like verbally speaking natural language. We have tools that you can do design without having any design skills. So, I think it's opened up a whole new site for agencies, consultancies, companies, but it's also opened a whole new site for a new like economy of like content creators. We had this first generation of social, we're gonna see another generation. And then underlying that is the infrastructure that the likes of Nvidia, Microsoft, Google, Meta are building these frameworks. I'm liking what I'm seeing about, they talk about interoperability, the ability to have agents talk to each other, the ability to have collaborations. I've never seen so many collaborations.

0:16:27.4 Michael Olaye: I remember I remember 2020 saying I've never seen so many tech companies talk to each other, like Microsoft, OpenAI, but OpenAI is also talking to Apple and they're all sharing the same... It's amazing. So I would say those are the two camps. Internal start now be AI ready, train, upscale. If you have a license platform you haven't renewed in 10 years, maybe now's the time to renew it so you get all the AI benefits. Then on the other side is governance to protect your organization. So that's the two bits.

0:16:54.9 David Chao: I fully agree with that. That matches everything that I'm seeing and hearing in the conversations I'm having as well, where data governance has really broadened to encompass AI governance. And absolutely your comments aren't biased, and privacy and IP infringement, I think are top of mind for all companies, and that's one reason why Alation has invested so much in AI governance as a solution. What are some best practices that you're seeing in terms of how to mitigate those governance risks that you just mentioned?

How do you mitigate the risks of AI?

0:17:21.3 Michael Olaye: I don't think we're at a place yet where we can trust the machines. Not even, I don't even say trust the machines 50%. If I'm being honest with you. I think we're in a place also where there's a lot of wishful thinking around what's responsible to like release to the world, what's responsible to train. I think we're all learning at the moment. So I think making sure that when you get unstructured data and you structure it, or where you have to structure data, that you clean it and you check for human error, you check for bias in the system. When you train the data, you train it on a diverse range of data that reduces bias. I'm not expecting us to eliminate bias. I think even in society we are biased. So machines are gonna pick up on that. When you build anything with AI, having a human in that loop where we are today, having humans in that loop, checking that also, it's good.

0:18:08.0 Michael Olaye: And I'm actually one of the people who truly believe like AI is more of a companion than a replacer. So I think even, yeah, of course maybe 20, 30, 40 years from now, as quantum computing comes online and all this other stuff that's coming. Yeah, of course AI is gonna be doing its own thing, we don't have to worry. But I think in the next five to 10 years as it advances, we still need humans to have those checks and balances. Especially if you're learning robotics and you have rogue humans in the world, you definitely need checks and balances. So I'd say in our industry it's understanding where we can use the data in the cleanest form and it's truest form, and then validating and vetting everything that we launch or push out that influences other people. And I'm not gonna go into like all the bad characters, but you know what I mean, all the things that happen and some of them happen by mistake. Not everything happens deliberately as well with these two.

How can people get started with AI? Tips and tools to try

0:18:57.0 David Chao: I think those are some great guidelines and north stars to aim towards. For folks that might be thinking, how do I get started? What should I be doing over the next month, the next quarter in my company so I can start this journey? What might be some practical suggestions that you might give?

0:19:12.6 Michael Olaye: If you are not using any of the tools right now, that's a good place to start. So many tools out there, I like to recommend the free ones, Leonardo AI is pretty good. If you wanna play around with AI video and imagery, please don't launch anything from it, make sure you do your due diligence, don't launch anything to the public. Waldo is a pretty good tool if you like to do strategic thinking. If you work for an organization that has, you know, like Microsoft, and Copilot is pretty amazing as well. If you work for an organization that has Adobe, Firefly is pretty amazing as, I mean, there's just so many tools out there.

0:19:44.6 Michael Olaye: If you work in the data field for like Salesforce or Optimizely or Contentful or ContentStack or all these different platforms, I think we would all struggle to find any enterprise or open source established platform across any entity that doesn't have a form of machine learning or artificial intelligence in it somewhere. So I would say start there. If you are in a more mature organization, it doesn't hurt to find a few other people that think like you to try and create a small crew where you can experiment.

0:20:13.7 Michael Olaye: I think experimentation is very undervalued in companies that make a lot of money or companies that are comfortable where they are. This technology allows us to take data that otherwise wouldn't really be used for anything and create new magic out of it. So, I can't say what that magic is because I don't work in those organizations, but these people we're talking to do. So, I would say try to find the magic or try to find the team members that can help you unlock the magic in your business to help build new things. Or if you just wanna play around for fun, create your own channel. Mess around, you know, share stuff internally. I think there's so many tools out there. Easiest way to start is obviously a ChatGPT is a good start and a little bit of AI is a good start.

0:20:52.0 David Chao: Well, thank you for sharing that.

The dangers of bad data in AI: Uber’s misstep

And you are starting to touch on the topic of bias and how data can be used for good, but it can also in some sense be used for bad as well, where there's the false truth being attached to that data. Is there an example that you can share maybe where data has not been used in the right way and has maybe led to poor decision making?

0:21:12.6 Michael Olaye: Yes, not on my personal experience 'cause we have due diligence and everything is checked. But there was the case of Uber in the UK. I can't remember what year it was. I think it was maybe 2017, 2018, don't count me in that. But they built an AI module into the validation of the drivers. And overnight, obviously, there was a lot of bias in the data. So overnight, a lot of, well, I say like Asian, like Pakistani, Indian, specifically men got blocked from the app. Like they couldn't drive the car. They couldn't do their jobs the next day because basically the app, when the data was trained, mapped them to be more criminals than actually drivers and actually just blocked a whole bunch of people, which took a lot of livelihoods away. And there was an uproar about that. So that is how bias can literally destroy someone's career.

0:21:57.6 Michael Olaye: The other super famous one that everyone has seen is the hand dryer or the soap dispenser. I don't know if you've seen that one, which is pretty famous online. Two friends, one's white, one's black. The white guy puts his hand under the soap dispenser, it dispenses the soap. And the black guy puts his hand on the soap dispenser, it doesn't dispense the soap. So again, you know, technology is great when it works, but when there's bias in it, it can be super, super harmful.

0:22:20.3 David Chao: And what are some best practices that you recommend to your clients to avoid that kind of bias going forwards?

Best practices for mitigating bias in AI models

0:22:26.2 Michael Olaye: I think it's multiple things. It's one, the team composition, like whoever's working on a project, whoever's working with data, having a diverse team where you can take all types of input is one. And even having that before you actually get the data, like looking out for what bias exists currently and how to find that within the data is super important. The second one is when you are training that data, make sure you have the right diversity on your team, basically, or the right diversity to check the output of that data and make sure it doesn't have bias. It's super hard to have any like large chunk of data without any bias in it. But what you wanna do is kind of reduce that by having checks and balances in place. And there's no better checks and balance than actually the actual people it's gonna offend in the end. If you have them on your team or you have people who lived experiences of biases, they're more likely to be able to flag it earlier on so that you can actually check for it in your products and your services.

0:23:20.3 David Chao: I think that's a great point. And I think having that diversity of thinking and approach and perspective is so important, not just in data, but in business and life as well.

0:23:29.7 Michael Olaye: Exactly.

How do you find high-value data? Michael’s approach

0:23:30.6 David Chao: I'd love to build on some of our earlier topics and talk a little bit about the data journeys that you've been seeing across the companies that you work with. At Alation, we talk about customers going on their own data culture journey, being able to find, understand and then trust the data sets that they work with. And I'd love to use this next part of our discussion to really focus on those themes. In your work, you're parachuted into these clients and you're obviously asked to get up and running very quickly. The first thing is, how do you even find the data sets that you need to inform your day-to-day work?

0:24:02.3 Michael Olaye: I think there's an art to listening. And I don't know if it's a specialty you build over experience or if it's a specialty that you have as you get good at your craft? But when you have a good team, 'cause I don't think it's any one person, but when you have a good team, there's an active way of listening and picking up on cues and picking up on conversations or phrases that the client might use. We know in the marketing sphere, sometimes you go to big organizations and brand is in charge of everything. Brand is the one that makes the decisions on where the marketing spend goes to. You go to other companies, that is the CIO, it's the IT team. You go to another company and it's actually the product team.

0:24:36.9 Michael Olaye: So being able to sit down, especially when you're in a complex client infrastructure and being able to, first of all, listen to who's there, but also try to spot who's not in the room. So if you hear a client talking about digital transformation, it's gonna be data-driven and it's all about creativity for data, but then there's no like IT person in the room. There's no CTO, there's no CIO. You tend to ask, okay, great. So we're gonna go into an audit phase or discovery phase, or we're gonna immerse ourselves in your business. As part of that journey, we look to a structure in the way we approach that to make sure we're speaking to the right people.

0:25:09.8 Michael Olaye: So I would say the art is in how you set up as an organization and how you spot the opportunities for not for you as a business to make money, but for the client to get the best services out of you. So that might sound too simple, but it's basically that some clients do not know that they are sitting on gold. Like they do not know that. They have tons of data that they've never done anything with. And then they kind of focus on the most simplistic things, media, SEO, social media content, website content. And then you come in and you're like, hey, we can help your customer service be more efficient by understanding how the data, how long it takes a call to go through. We can help you process products more better by understanding the transaction from seeing something online to going in-store to buy it, it's retail, you know, looking at those data sets and seeing patterns or bringing them together to see journeys. That's kind of where the secret lies.

0:25:57.0 David Chao: And you mentioned many companies are sitting on these seams of gold, this treasure trove of data that they might not be leveraging in the right way. Is there an example that you can share where that's been the case?

0:26:08.1 Michael Olaye: I think any organization, I'm not gonna name any particular client, but I think any organization, maybe not now, but I'd say maybe around the 2014, 2015, 2016 era, I think any brand that was like 50 to 100 years old, didn't work with any external partners, built everything themselves, had legacy systems that haven't been updated for 20, 30 years, they tend to sit in a lot of data. Because people only tell the stories they need to tell with that data for the benefit of themselves. So if I'm building new products, I just tell the product part, the data part for the product.

0:26:37.8 Michael Olaye: I don't need to worry about the whole organization. But I think as the internet's evolved, AI's come on board, people are now looking at how you have got all this data. We had big data happening all 2016, 2017 before AI went crazy. I think a lot of brands took a step back and a lot of partners were asking those questions like, hey, you can build unique things that put you at the front of the market if you have X, Y, Z. And they were like, hmm, we don't know. We have all this stuff we've been collecting for the last 50 years. Let's have a look. Give an example. There's a brand called Aegeus in the UK, I think they're global actually. And they came and said, hey, we have this unique opportunity to service students with a different type of insurance, right?

0:27:16.2 Michael Olaye: And as we were talking through that, I mean, the data, the research they've done was amazing. They had worked out that students never need to insure everything at one time. They only need to insure certain things at certain times. So they actually built a whole business unit and a whole business plan around elastic insurance, which is the ability for any student when you're living in the house to insure three things. So you can insure your laptop, your keys, your bike, and then when you come home, you can switch it. You can just open an app and go then change it to some three other objects. So if you're going out for the night, you just have your wallet, your phone, you can switch it to two objects. And I think that's the really data-driven, really cool insight to actually how behaviors occur and how you as a business can play in those behaviors, as opposed to trying to convince people to buy your package that you've created. They actually use the behavior to create this whole new business unit through data.

What holds companies back from innovating?

0:28:00.9 David Chao: That's a great example of a company using data to drive innovation. What do you think's holding the majority of companies back from following suit in that way?

0:28:08.7 Michael Olaye: I wanna say a lot of companies stumble on innovation by mistake. I think organizations work in an exploit manner, which is basically you've spent a lot of money, you've built something, service or product. It's doing well. Everyone doesn't get fired for being safe. So everybody plays safe. And then you might have an R&D division, or you might have a group of people who go off and do a moonshot project, or there's an R&D department who's experimenting with a tech. That tech might then gain traction and somebody might try it on the business and then it goes good and then it's called innovation. But then you have, I think, the more mature companies who are forever evolving, not because they have to, but 'cause it's a necessity to stay as a market leader.

0:28:47.7 Michael Olaye: I think those organizations are very intentional on how they invest to use the data they have, to use the industry knowledge they have, the experience they have, to look at new ways of either evolving the products and services they have or building new ones. So for me, innovation sits in like this weird thing where you have so many different types of innovation. You know, you have frugal innovation, you have creative innovation, and it's just understanding which organization and how they play. Are they more of an exploit organization where, you know, if you wanna innovate, then it's gonna be something small. Or are they more of like an explore organization where they actually, they always trying out new things. And you can see that with Google, for instance, you know, they kill more projects than they keep alive. I think from all the different things they build on the Google platform, there's a website called the Google Graveyard.

0:29:32.6 Michael Olaye: I don't know if you've ever heard of it, you should go check it out. It's all the Google projects that have ever been killed. There's a lot on there. We all remember Google Plus. Remember, you've been plugged into YouTube and it didn't really like work as they try to force you to use it. So I think innovation is not always successful, but I think it's always necessary.

0:29:49.5 David Chao: Agreed. And that website sounds super interesting. I will definitely have a look after this. Let's go back to this theme of finding, understanding and trusting data. We talked about the find part. When we talk to our customers, oftentimes understanding is also a big blocker as well, because different teams have different definitions of how they think about, for example, what is a new customer or even how they think about the different parts of a customer journey. Is that something that you've seen in the clients that you've worked with?

Understanding data and the brand-loyalty debate

0:30:17.0 Michael Olaye: No, I actually think the clients that I've been lucky enough to work with, they've always really known who the customer is. The challenge hasn't been that. The challenge has been how to keep away the new like challengers who build really cool stuff or who offer really enticing services. And I think something we don't hear much around now is like brand, like what is the role play? So I remember, you know, 2010 to 2015, it's a lot of conversation around. I spent most of my time in the UK, I moved to the US two years ago. So, a lot of the conversation was around is brand important? And there was two schools to that. Some people said, nobody remembers your brand, nobody cares about your brands.

0:30:57.3 Michael Olaye: People today can switch brands really easily. And then there's another school that would say, actually, no, like if you use the products, the data, the information and the experiences that you build for consumers and you build like really fresh stuff, people will love the product, which is technically they also love the brand. And we know the Amazons and the Ubers and the Googles and the iPhones versus Android, all those conversations. So I think many brands know their customer, where they do struggle is like, what's the next best customer? That's kind of where the issues usually arise. It's like, okay, mama and papa have been with us 20 years, but the children have no interest in buying our products. How do we get them? Like how do we bring them into the brand? How do we show them the products we have? How do we service them in the right way? How do we stop them leaving? Because obviously this generation has a lot more options than 20, 30 years ago.

0:31:47.9 Michael Olaye: And if you serve as someone, you can spend all the money you want on TV and advertising, but if you build a bad product or you build a bad experience or your customer service doesn't work, or you can't see the complete user, then you could easily lose them just as quickly as you gain them. And that's just a waste of money. So, a lot of the clients that I've been lucky enough to work with, even now, it's not so much like who the customer is now, they know and they're building and adapting to suit that customer. But it's also building a business that's sustainable for the future. So they try to work out like who's the next best customer. Getting that single customer view with all the different touch points, attribution, tagging of the data, using the data in ethical ways and things like that. That's kind of where I think a lot of businesses are focusing now on the things they're building. Of course, I've got a skewed vision 'cause I sit very much on the digital product side. So that's what I've been seeing the last couple of years.

0:32:35.6 David Chao: And you mentioned there's these two schools of thoughts, brands that can engender loyalty on an ongoing basis. And then another school of thought, which is saying, hey, customers don't have brand loyalty and are gonna switch based on what best serves their needs in that moment. Which camp do you sit in?

0:32:50.2 Michael Olaye: I sit in the camp that people do have a loyalty to brands. I truly believe when you interact with a brand the first time or second time, you build an affiliation or hatred for that brand. That's just my opinion. It could be anything from the brand you choose to go do your shopping in. I will drive past one supermarket to go to another one, not because I think the other supermarket is worse, but I have a comfortable feeling of what I'm gonna expect of my experience in that store. So, I'm gonna go to that store. And it's always that first leap of faith. What is the thing that made me pick that store for the first one? It could be a variety of things. It could be I searched for what times they were open and it just happened to be open later. Or actually when I searched for their website, I actually got a lot more information about who they are and what they sell than the other brands. It's that input or I saw a commercial and it resonated with me personally.

0:33:38.8 Michael Olaye: But then I've also had brands where I've used them once and then thought never again, because you stay with a brand for five, six, seven years and you think they're treating you bad 'cause they're taking a bit longer than a minute to respond in a chatbot. So you try another brand and you see just what true terrible customer experience is and you go straight back to the brand. So I do think there's an affiliation. And I think with all the information we have out there, we humans are super bombarded with like just tons of it.

0:34:03.5 Michael Olaye: Like think of all the data we as humans have to process every day. Like it's just stuff coming from everywhere. So when you find the right brand, when you find a brand that has values like you or is ethical or has great customer service or builds amazing things or products, I believe people stick with it. They don't like to admit it. Look at Ben and Jerry's, Nike, Adidas, Netflix. I mean, the list is endless. Uber, DoorDash, there's all these amazing brands and they're not all that old, but we all know them and we all use them. Like people queue up for iPhones even though they look like a last year's iPhone, but they have a small tweak internally. People queue up for them anyway. That is brand affiliation. That's brand love. I don't think they're buying it just because the tech is the best.

The future of AI: Technology stacking

0:34:43.1 David Chao: Let me ask you one last question. You've been so generous with your time and I've really been just so inspired by the journey that you shared going door to door in Soho in London, trying to get in and eventually seeing vector graphics on Flash. To fast forward now, we're talking about AI generated images as a business and marketing and creative leader in Los Angeles. That journey has really come very, very far. When you think about this next stage of your journey, what are you most excited about next?

0:35:13.8 Michael Olaye: So I've always been a disruptor, right? So I'm excited about what I call technology stacking. So right now we have a lot of these AI tools, because it's natural when you create a tool, you wanna monetize it, you wanna grow your business from it. But I think we're going into a world where the automation is gonna be so advanced and the understanding of how data protection works will be so advanced and hopefully there'll be regulations to control that. I know we didn't really cover regulations with governments and stuff, but that's coming as well. And we'll need these agents, you know, the learning agents, which we have now, you know, some of the chatbots are more linear, they kind of do one task, but the learning agents, like we saw the first jump of that.

0:35:49.6 Michael Olaye: We saw the Rabbit R1 come out and we saw the AI pin come out and everyone laughed at it. I didn't laugh at it, I thought it was amazing. It just reminds me of the first smartphones that came out with Palm Pilots and stuff and then they died and then the iPhone came. So, these tools we've seen with this amount of data and the things we're seeing with AR and blockchain and all that stuff, I'm excited because we've seen the first round of that, you know, the hype phase has come and gone. The next round of technologies, when the iPhones and the Androids from factory have what I call multi-modal models in them, like literally you can talk to your phone and say, Hey, I've just finished eating dinner, how do I get home? And you won't have to open the Uber app. You would have to open your screen and go to the Uber app and tick that and put your credit card. The agent will do it for you. It will know, right, do I go to the Lyft app or the Uber app?

0:36:38.0 Michael Olaye: Do I have to turn the heating on at home before Michael gets there? Because it's gonna take 20 minutes for the drive. These agents that are being worked on now, which I call, you know, like the learning agents really, that's what in the medium and the long term is coming. So, you know, multi-agent systems, AI systems stacked on top of each other, AI systems collaborating across and we, the users having more control of our data, it's gonna be super exciting. So yeah. So I am on the lookout for like, what's those next hardware that is more affordable. Software is a whole different conversation 'cause the AI can probably write the software itself. So you can imagine the acceleration of like growth of code. So when that happens, when the AI is writing the code and the hardware is cheaper and we have this infrastructure, like that's gonna be super, super, super exciting.

0:37:25.1 Michael Olaye: I think I was a big IOT fan back when, and that also went a bit quiet. And I think we're back in that phase with just the current IOT. 'Cause I think gadgets, buildings, systems, data centers, products, apps, cars, they'll all talk to each other and we'll all have agents, whether we like or not. And they will all contain our data and they will manage it for us. And people would have to earn the use to access that data and have to earn the use for us to wanna stay with them. So, I think I'm in the industry that will be at the forefront of some of that. So I'm pretty excited for what the future can bring.

0:37:57.0 David Chao: I am as well. It's certainly an exciting time to be in this space. And thank you again, Michael, for sharing these insights and these stories with us. It was super informative and excited to explore this future together and see where this industry and this space goes from here.

0:38:12.2 Michael Olaye: Thank you so much.

0:38:16.2 David Chao: What a fascinating conversation. Michael's journey illustrates the power of harnessing data for strategic decision-making and innovation. From using AI to personalize customer interactions to overcoming biases in data collection, Michael's insights underscore the importance of a diverse and thoughtful approach to AI. Michael shared his vision for a future where AI makes businesses more efficient and helps brands connect better with customers. He also highlighted the need for strong AI governance as businesses embrace AI for digital transformation. I'm David Chao, CMO of Alation. Data Radicals, keep learning and sharing. Until next time.

0:38:51.4 Producer: This podcast is brought to you by Alation. Your boss may be AI ready, but is your data. Learn how to prepare your data for a range of AI use cases. This white paper will show you how to build an AI success strategy and avoid common pitfalls. Visit Alation.com/ai-ready. That's Alation.com/ai-ready.