Pod Insights The Brand Called You Mastering AI Beyond the Hype | Cecilia Dones, Founder, Chief Data Officer, 3 Standard Deviations About this episode In this episode of The Brand Called You, Dr. Cecilia Dones, data leader and AI ethics expert, on moving businesses from mystical AI hype to mastery — leveraging data responsibly, balancing automation with human touch, and empowering organizations to innovate ethically.00:33- About Cecilia DonesCecilia is the founder and Chief Data Officer of 3 Standard Deviations.
Transcript 0:08 Speaker 00 Welcome to another episode of The Brand Called You, a podcast and podcast show that brings you leadership lessons, knowledge, experience and wisdom from thousands of successful individuals from around the world. I'm your host, Ashwathosh Garg, and today I'm delighted to welcome a very accomplished professional from New York City, USA, Cecilia Doness. Cecilia, welcome to the show.
0:31 Speaker 01 Thank you so much. I'm excited to be here.
0:33 Speaker 00 Thank you. Cecilia is the founder and chief data officer of Three Standard Deviations. So Cecilia, let's start by asking you, tell me a little bit about your own journey in brief.
0:46 Speaker 01 Yeah, it's fascinating in the sense that it was completely nonlinear. So what has been my north star in everything I've ever done is to follow my curiosity. And so I am a qual-quant researcher, very much interested in telling stories about people using data. And so every opportunity I've had, every success, every failure has been, I'm curious about that. Let me understand that better. Is there something more to that idea that I can explore that helps me understand the human experience? And so the primary utilization of my skills have been mostly in consumer spaces, so marketing and advertising.
1:29 Speaker 01 I would say the second part of my career, I got really excited and wanted to do more with the data, more with the analytics. I'm a trained statistician, so that's the quantitative side. And I kept running into the same problems. My data is over here, my data is over there, I don't have access to my data, nobody trusts my data. And so I started picking up more and more data skill sets. And so within the last 10 years or so, I've been brought in as executive leaders into organizations who are trying to build these data and analytical capabilities inside of their organization. So I became more of an enterprise data leader. And so worried about marketing, of course, but also worried about finance and operations
2:10 Speaker 01 and supply chain and in sales and all of those other functions within an organization. So really building at bigger scales across all touchpoints inside of an organization where you're trying to make decisions.
2:23 Speaker 00 Absolutely.
2:23 Speaker 01 And then the last part, I would say My pandemic gift myself was to attempt to finish my doctoral dissertation. And I will say, as of a few weeks ago, I passed my orals and so I'm officially doctor now. And I've been teaching AI ethics. I have been exploring the applications of AI, generative AI, agentic AI inside of organizations across functions. And it has been a wonderful journey. Still trying to figure out why people do the things they do.
2:58 Speaker 00 I know. Thank you so much. And congratulations, Dr. Cecilia Dones.
3:05 Speaker 00 So, you know, when I was reading about you, I was fascinated to read that you talk about moving businesses from mysticism and magic to mastery.
3:16 Speaker 01 Yes.
3:17 Speaker 00 Can you break down what this transition looks like?
3:21 Speaker 01 Uh huh. So, the reason why I say mysticism to mastery is for many organizations, regardless of the maturity of that organization, how much I use AI in my day to day business, it still seems magical. We don't understand what's inside the black box. Somehow I type some things into the machine and something comes out that may be meaningful or may be valuable to my organization. So it feels a little bit like magic. My reasoning for how I want organizations to mature through this process is that, yes, it can feel magical and it can feel exciting and all of those things. However, if we really want to fully utilize and realize the value
4:06 Speaker 01 from our data and analytical practices, we have to master it. And so what that means is, are we actually utilizing data and analytics and AI to change business outcomes? So it's not enough to say, I did the analysis. It's not enough to say, oh, I put the data here or there. What are you changing in the way you're making decisions inside of your organization such that the outcomes of your business is changed as a result of AI? And so that slight nuance from focusing on outcomes as opposed to output is the distinction I make when you go from mysticism to mastery.
4:43 Speaker 00 Great response. Thank you. Cecilia, many AI models function as black boxes. How can businesses make AI more explainable and trustworthy for their customers?
4:57 Speaker 01 I am a big fan, and as I mentioned, I do teach AI ethics. I'm a big fan of responsible AI. I nuance that definition a little bit. So explainable, I think it's great. I think some of the more reasoning models that are more lately coming to date is helpful. However, I think explainability doesn't, it's not sufficient from a consumer standpoint. Meaning, I can say all of the words. However, if you, as the recipient of that message, doesn't understand me, it doesn't matter that I'm using all the words, all the explainables. No, I press organizations to focus on what is understandable? What is expected from your consumer?
5:43 Speaker 01 And so if you have a consumer that is interested in, I don't know, flavors of cake, that's top of mind at the moment.
5:53 Speaker 01 And you have a machine that's helping you to offer up different variants of cake frostings and cake flavors and all of those things. What would be important for your consumers to understand? Oh, how did the machine decide that chocolate and cherry goes together? Oh, maybe that's important to your consumer in that context. As opposed to if you were to explain chocolate and cherry in maybe a financial services context, that would be confusing to a consumer. And so I think it's very important for organizations who are trying to remove the opacity around these models and decisioning process is to think of it from the consumer's perspective. And what is their expectation in the transaction? If I'm buying a cake, maybe you should know about flavors.
6:37 Speaker 01 If I'm trying to open a bank account, maybe you should know about the variable rates on loans and lending. So it's really contextually driven.
6:45 Speaker 00 Very well said. Your work also focuses on how technology mediates the lived experience for individuals and brands. Can you share an example of how this manifests itself in everyday life?
6:59 Speaker 01 Yes, so it's related to my doctoral research. I was very much interested in understanding nonverbal trust and authenticity signals in technology mediated communications in virtual environments. And it's a very fancy way of saying, how do we talk to each other? How do we listen to each other when we have the machines helping us to do that, especially online? And an example of that is literally right now, this podcast, this videocast. We're utilizing technology, we're using the internet, we're using Zoom to be able to connect to each other. In another lifetime, in different times, maybe it would only be through the telephone, or it wouldn't be at all, or maybe only over letters. It's
7:40 Speaker 01 very different. So how do we interact now, knowing that we have machines that help us to see, we have machines to help us to hear, we have machines that are helping to translate our language and our communication over many, many miles, many, many distances?
7:54 Speaker 00 Yeah, well said. And given the rise of deep fakes and synthetic data, how can consumers differentiate between genuine and AI-generated interactions?
8:07 Speaker 01 The short answer, and maybe it's a bleak answer, is no, we cannot tell.
8:13 Speaker 00 We cannot.
8:14 Speaker 01 We are past the point where we are able to meaningfully distinguish. And of course, there's very extreme cases. So for example, if you have an image and there's, I don't know, seven fingers as opposed to five fingers, and they're all twisted the wrong way. Yes, there's the obvious cases. But in for most AI-generated, most average individuals will not be able to do, be able to tell. So what do you do in this case? I would say, instead of fixating on, is this real or is this not real? I would ask myself, what is the incentive for me to believe that this is real? If I were the other party, if I were the other person, why would they want me to believe this is
8:58 Speaker 01 real? Do they get any additional benefit that could be at my detriment as a result of me believing that it's real? And understanding those incentives and motivations can help me critically think through, okay, this image, or this audio, or this text I'm reading, how much do I want to trust it? And I think that's a more pragmatic way of addressing is this real or not? Because at this point, it's very, very difficult to do so. We don't have enough technologies at this moment to really address authentication, especially in these types of environments.
9:33 Speaker 00 Wow. That is, in one way, a good reason for worry for an average user like me and maybe something to look forward to as well. But Cecilia, you've also written about synthetic data in marketing. How can businesses leverage it without compromising authenticity?
9:53 Speaker 01 It's very, very difficult. And there are many use cases for synthetic data, especially in marketing, marketing research, in those sub disciplines. My main advice for any organization and I do advise smaller organizations that are trying to enter this ecosystem providing services is that how are you going to determine that the data that it is producing is representative of the data you want it to represent? So is there a truth set? Does it follow the distribution of the underlying data?
10:31 Speaker 01 Are there hidden biases that you're codifying as a result of creating synthetic data models on top of already biased data sets? And so being very cognizant of the potential for misrepresentation is something that I really advocate for anybody who's utilizing it. However, can it help us with imputation? Oh, yes. If there's missing holes in your data, yes, it can be quite helpful. Can it help speed up research? Oh, yes, it can definitely create efficiencies when you're trying to do things at scale. Can it help us with exploratory? So I don't have a lot of resources, but I kind of want a direction for my research to go. Can synthetic data help you?
11:16 Speaker 01 Oh, yes. Is it the final end all be all?
11:19 Speaker 00 Yeah. Great response. Thank you. And I must compliment you. You explain these things so easily and simply. I'm sure my viewers and listeners will really love every single comment that you're giving to us. My next question, Cecilia, is with the rise of generative AI, are we moving towards a future where human creativity is augmented or replaced?
11:44 Speaker 01 I am an optimist when it comes to these things.
11:48 Speaker 00 Yeah.
11:49 Speaker 01 I think for any attribute that we would describe as humans, so creativity, critical thinking, all of those terms. I would say AI is going to redefine it. It's going to force us to reflect and say, is that really the definition? Is that a meaningful definition today? And so I am firmly of the camp that creativity is going to be augmented, not necessarily replaced.
12:16 Speaker 00 Hmm, thank you. My next question is that how does AI impact decision making at a subconscious level for customers?
12:28 Speaker 01 So the question, this happens a lot with recommendation systems in particular, and we've had recommendation systems for a very long time.
12:37 Speaker 00 Yeah.
12:37 Speaker 01 And so when we think about things from a subconscious perspective, when I say I like something, or when I say I want a particular product or service, the question I ask myself is, had I not been recommended by the algorithm, if I had heard it from a friend, would I still like this item? Would I still be seeking this service? A book that I'm actually currently reading, it's relatively older, is You May Also Like by Tom Vanderbilt. It explores these ideas of wants and likes. And when it is influenced by algorithms, what does that mean? So how do we be true to our own wants and likes, knowing that the machines are kind of helping us suggesting things?
13:20 Speaker 00 Right, well said. Many businesses implement AI-driven automation, but customers still value human interaction. How can brands strike the right balance?
13:36 Speaker 01 Whenever there's a question regarding what do we do for the consumer? I love the question because your thinking, your strategy is in the right place. I always default to the same answer. Let's go ask some consumers. Depending on the sector you're in and depending on the maturity of the relationship you have with your consumer. Meaning, is this purely a utilitarian, transactional type of relationship with the consumer? Or is it more transcendent? Meaning, the consumer has, sees the brand as part of that identity. So in luxury, for example, many consumers will actually identify with the brand.
14:15 Speaker 01 Oh, I am a Prada user, I am a Gucci user, that's why I have this product. As opposed to, I don't know, maybe cereal or milk or other goods, which may be a bit more utilitarian. And so, I always say the same thing. You have to ask your consumer, what's most appropriate for you? So if you're in a utilitarian category and maybe it's only about replenishment? Maybe AI-driven automation is excellent because you're on a replenishment cycle. You don't need too much interaction with humans. However, maybe if you're in the luxury sector and you're buying a really, really expensive handbag?
14:56 Speaker 01 You like that personal experience that the concierge knows your name, knows your family's background? Knows the exact thing you bought last time, your last holiday, all those things. And so, understanding the sector you're in and understanding the mindset of your consumer, of what they expect from you? Will help guide what's appropriate to augment with AI? What's appropriate to replace with AI? And what really should be human-to-human interaction?
15:23 Speaker 00 Amazing. My next question, Cecilia, is how should businesses upskill their workforce? To work with AI rather than fear AI?
15:34 Speaker 01 So, culture change is always difficult, so I want to acknowledge that. As leaders, there are a couple of components, I think, that's critically important for culture. Creating opportunities for psychological safety, because you had mentioned fear. When people are afraid, they stop learning. When people are afraid, they start behaving in defensive ways. And so, that doesn't help at all with trying to create AI adoption. It's important to acknowledge it, but as leaders, how do you create the safe space? What I mean by that? Are you creating forums where, I love this example, I've actually implemented this a few times in different organizations.
16:18 Speaker 01 Do you have a laboratory where you just play with the tools? And if there's an outcome, great, you have a new potential business idea and you can implement it into your line of business. Excellent. Or maybe not, and it's just about them just learning how to utilize the tool. Are you creating those spaces for people to actually play with the tools? And then, everything related to AI, you cannot read it in the textbook. You cannot talk about it. It is an experiential thing. If you want to be an expert, if you want to be skilled, you actually have to play with it.
16:54 Speaker 00 Amazing. My next question, and you're a data scientist. Do you think that at some stage, AI will eventually start to replace a lot of data analysts? Or will it simply change their roles? And please excuse me if I'm offending you on this question.
17:15 Speaker 01 First of all, curiosity, I'm never offended by curiosity.
17:19 Speaker 00 Okay, thank you.
17:20 Speaker 01 So, thank you. And asking the hard questions, we need to be able to do that. So, I appreciate that very much. Will AI replace data analysts in the form they are today? Yes. And why I say that is that, as I alluded to in the previous... Question, AI is giving us the mirror to ourselves to reflect upon is are we utilizing the right definitions for creativity? Are we utilizing the right job descriptions to describe data analysts? So for example, what I believe will happen for data analyst roles is that AI will augment what they do in terms of repetitive tasks.
18:02 Speaker 01 However, where data analysts will become more skilled is the contextual components. So really understanding the data generating process that goes into these data and analytical tools. I find especially in many organizations, we do not spend enough time understanding what do these ones and zeros mean? Where did they really come from? What is it supposed to represent?
18:26 Speaker 00 Yeah.
18:27 Speaker 01 We just look for the ones and zeros. And so if machines are dealing with the ones and zeros for us, we can start to say, oh, was that the right measurement tool that we should have used to collect that data? Oh, wait, we don't really understand the provenance of this data. Oh, maybe we should not trust it, or maybe we should test it further so we can determine whether we trust it. It's going to redefine the data analyst roles such that they're going to be elevated to do more critical thinking and more creativity with the data. So I get quite excited. In the near term, is there going to be disruption? Yes, that's always what happens whenever you have exogenous shocks into an ecosystem. Will a lot of people be displaced? Yes. Will a lot of people feel lost?
19:12 Speaker 01 Yes. Will a lot of people get creative and find new ways to create value? Also, yes.
19:18 Speaker 00 Yes. What a great response. Thank you. I have time for three more questions for you, Cecilia. My next question is, how can smaller businesses, which may lack AI expertise, still leverage AI for growth without relying solely on big tech platforms?
19:39 Speaker 01 I think when you are resource constrained, and it's exciting in the sense that if you're resource constrained, you get to be a little bit more creative because you have less resources. So maybe smaller players who are trying to enter the ecosystem, you suddenly become very exciting because now you're both trying to grow and that could be potential opportunities for piloting, piloting and innovating things that are very, very new in the ecosystem. However, I would say if you are trying to figure out, how do I truly embed this? I think you have to be much more mindful about the use cases you choose to pursue.
20:20 Speaker 01 And what I mean by that is, what is the core of your business? I always start there. What is your business strategy? And then from there, go to your data strategy, go to your analytical strategy, go to your AI strategy. If your business strategy is not clear, all the technologies in the world will not necessarily help. There will still not be a product market fit. And so I start with a business strategy first. And if we go through that exercise and we determine the data strategies, right? The analytical strategies, right? And then you need AI? Then it could be quite exciting to utilize small players. So it's not to close the door, but you have to be a bit more strategic and mindful of the choices you make, especially when you're resource constrained.
21:06 Speaker 00 Thank you.
21:08 Speaker 00 If you were to change one thing about how businesses currently approach AI, what would it be?
21:19 Speaker 01 I would say being more mindful of what's hype.
21:29 Speaker 01 It's exciting. The, the, the velocity of change that's occurring from a technological standpoint. So I get it. I get excited too. However, who are your customers at the end of the day? Are they machines or are they people? Last time I checked for many organizations, it's people. And so. I would always try to ground myself in how am I actually servicing my customer, my consumer? Are they expecting faster, better, stronger, all the things? Before I start thinking about, oh gosh, should I AI this? AI is not a verb.
22:10 Speaker 01 And so I would be very cautious about getting overly excited about just implementing technology when there really isn't a value proposition that's consumer facing. And so I wish organizations would be more mindful of their customers and consumers at the end of the day and then decide whether or not AI is appropriate for that value delivery that they're trying to bring.
22:34 Speaker 00 Very well said. And my last question to you, what is one piece of advice you would give to businesses looking to ethically and effectively implement AI in their customer strategy?
22:50 Speaker 01 I've spoken about this, I've taught this. No matter how many technologies we end up learning how to use embedding into our business processes, I am a big fan of human in the loop. If you're facing a customer, if you're facing a consumer and you're trying to build a technology experience for them, make sure humans are in the loop and make sure humans that represent your customers are in the loop. If everyone looks the same, I don't think your customers all look the same. So making sure you have that inclusion and accessibility when you're trying to build experiences for consumers and making sure that people are involved.
23:32 Speaker 00 Amazing. And on that note, Cecilia, I just want to say, thank you so much for speaking to me about your own journey. I think we covered some incredible ground on AI and how AI is increasingly becoming an inherent part of customer strategy. We covered a lot of very interesting points and I think I learned many, many new things from you about customer retention, about customers and the use of AI. Thank you for speaking to me and good luck to you.
24:04 Speaker 01 Thank you so much. This is great.
24:06 Speaker 00 Thank you.
24:09 Speaker 00 Thank you for listening to the brand called you video cast.
24:12 Speaker 02 Podcast, a platform that brings you knowledge, experience and wisdom of hundreds of successful individuals from around the world. Do visit our website www.tbcy.in to watch and listen to the stories of many more individuals. You can also follow us on YouTube, Facebook, Instagram and Twitter. Just search for the brand called you.
Previous The Hidden Architect of Wi-Fi | Cees Links, CEO, Father of WiFi; CEO, SuperLight Photonics Next Authentic Leadership Under Pressure | Charles Browne, Certified No more Mr. Nice Guy Coach; Author, 'The Pursuit of Identity & Purpose' Page updated November 15, 2026 · transcript and summary are machine-generated