How Agentic AI is Reshaping Consumer Relationships

Speaker A: M this is the Real Digital Transformation podcast series, empowering technology and business professionals to succeed with digital transformation. Now here's your host, best selling author and series editor of the Pearson Digital Enterprise Series, Thomas Earle.

Speaker B: Hi, this is Thomas Earle and welcome to the Real AI and Digital Transformation podcast series. Today I have with me Lev Mason, the CEO and co founder of Ask youk Target Market, also known as aytm.com. welcome Lev.

Speaker C: Thank you. Hi Thomas. Thanks for having me.

Speaker B: So we're here today to talk about the impact of AI, specifically agentic AI in the consumer enterprise strategies that organizations have been using for cultivating and growing their relationships with consumers. Um, and you are here to explain to us the three waves of the consumer insights that have been reshaping those enterprise strategies and how all that now has been culminating into the involvement of agentic AI.

Speaker C: Yes, uh, and it's something that I've been focusing on most of my waking hours for the past couple uh, decades and I uh, have seen and we continue observing these tectonic shifts and major uh, waves. And the first wave that um, has been around for many decades is the reliance on full service agencies for full service outsourcing of very high caliber specialists coming into enterprises and answering their consumer insights questions, whether it's about strategy of uh, their entire company expansion or more tactical ones related to pricing and packaging and optimization concept tests and uh, all others. But the way of getting to those insights was rooted in companies outsourcing that to uh, the full service agencies. And it came with all the great assurances of quality like you don't get fired for hiring IBM and uh, red carpet treatment and all other benefits inherent to that way of working. But it also was inherently slow, an expensive process. So rolling uh, back to 2008, 2009, there were a bunch of full service agencies, there were panel companies and there were a number of technology companies but all of those were not talking to each other. Right. It was quite fragmented infrastructure in our industry. And the wave of DIY do it yourself consumer insights started right about that time when ytm, um, being a small startup introduced a uh, novel idea that anyone can formulate their questions, program them lunch with a credit card and get answers a few hours later from their target audience without talking to salespeople. And that grew in complexity, that grew in popularity and uh, a decade later only uh, a lazy company didn't introduce some or all components of that business model. Uh, specifically panel companies moved and added survey tools and consultancy, uh, arms consultant companies also uh, connected their API, panel partners or had their own panels and added that DIY layer. And the technology companies also often started uh selling the full cycle and that created efficiencies that created uh speed to insights that uh, created a lot of uh understanding and knowledge of how exactly those insights were produced. Because with full service it's um based on solely on trust and uh on reputation of the provider. The second wave served uh us very well and many wise uh CPG companies and uh global consumer brands uh repositioned their talents to adopt this idea of hands on the keyboard and uh used different growing population of wonderful DIY platforms. Before we move to AIY wave I uh would like to provide a few examples of what it felt like in the full service and in DIY world in full service. Thinking back in the beginning of the century if you had a project you were looking at at least $30,000 to start a conversation, to sign a contract, to speak to several salespeople or a little army of very smart statisticians and researchers to be involved in this process to set up the parameters of what you want to get uh and um, implement all the smart methodologies on the back end and connect different companies as I mentioned a little bit ago and at the end produce a ah report that you would be um getting to advise your internal decisions. With DIY that timeframe and that expense uh significantly was reduced. Instead of waiting three months you uh were able to do similar job in a few days. And um, you still could uh use the very sophisticated models and research tests such as Maxdiff and choice based conjoint and orchestrate them within a particular platform that you were using and uh get access to um, general population or specialized panels around the world. Uh it felt much faster, it feels uh empowering but you still need to understand what you're doing because you can mess it up. You need to uh use uh, all those tools wisely and uh have the um experience and the research expertise and in order to achieve great results um and take advantage of the efficiencies.

Speaker B: Super, thank you. Go ahead.

Speaker C: And uh, the new wave is uh even more fascinating. It inherits all the benefits of do it yourself wave but it reduces the pre existing reliance on experience of the user. Because with AI and all the previous technology that companies uh were building for the first time, we technologists and market researchers uh joining our forces can encompass all the knowledge and methodologies and know how into a system that is there to serve a much wider spectrum of people with uh, helping them to avoid rookie mistakes, with understanding their natural language, with Translating what they want to learn into how to learn that properly and unpacking the insights, starting with the very classical statistical analysis and then uh, explaining it in the reporting visualizations and uh, the language that those marketers and brand managers and strategists and client success, employee success and uh, C Suite and other people, uh, the end users of the wisdom of consumer insights are actually seeking.

Speaker B: Super. And so, um, how then would the example you provided before be impacted by

Speaker C: the involvement of that uh, the real impact and real speed of it is yet to be seen. We are just at the beginning of the aiy, uh wave. But uh, the reality is that you will see results at very comparable speed, uh, as diy, uh, perhaps sometimes even shorter. Although the um, results of a quantitative study that are produced uh, in less than 24 hours sometimes are critiqued because uh, it's hard to make them representative. But the world of delivery from start to finish shrunk to hours. And the biggest impact of course is who can initiate it, who can benefit it. Uh, the speed is in not waiting for specialists available for you internally or externally, but interacting with AI agents directly to answer your questions while avoiding the problem of asking ChatGPT or any other LLM and suffering with hallucinations and uh, prediction of the next word rather than methodical consumer insights. Proper methodology.

Speaker B: Super. So, um, based on what you've seen so far, involving agentic AI in these processes and basically redefining how enterprises relate to their customers, to the consumers of what they provide, um, have you seen any challenges or obstacles or significant problems that have arisen from um, attempting to involve agentic AI in the automation and in enhancing the intelligence or the responsiveness of organizations? Is there anything that you've seen that you can let the listeners know that they should be looking out for?

Speaker C: Sure, sure. So first of all it's challenging to see uh, what is real gold and what just looks like gold. Right. We all have experienced that by now that we see a very convincing answer from AI and uh, we run with it, uh, if we're in a rush only to discover that uh, we made uh, a false claim in a conversation, uh, hopefully at a more informal conversation rather than at your formal um, board, uh, presentation or something. Uh, LLMs inherently are prone to um, justifying any set of uh, uh, uh, data and outcomes that uh, seem in the moment as most logical, uh, to predict, uh, based on the context that it was given. They're not evil, they're not bad or good. They're just, that's how they work and for non researchers especially, uh, it's the biggest challenge of the day I think is how to tell the difference between an output from one system versus an output from another system that uh, are equally convincing and they sound smart, they sound uh, like plausible and uh, the temptation to just take them on the face value and run with them is very high. The more technical uh, answer to this question that I think uh, will relate to a lot of companies like AYTM who are in this business trying to solve it is that LLMs that AI of this generation is focusing on are very good at summarizing and working with unstructured data. And that's why we see a generation of startups and uh, pretty well funded young uh companies focusing on qualitative research, qualitative being made of words primarily and conversations and interviews because it's much easier to structure them, to moderate them, to summarize and analyze them. The biggest challenge I think that uh, our industry is currently has a privilege and pleasure solving is how to marry the strength of the AI agents and LLMs with the quality quantitative data and classical statistical analysis of it without dropping the bar on quality, repeatability and performance of that study.

Speaker B: Right. And so um, how should organizations then approach the involvement of LLMs into their UM processes and workflows? Uh, to what extent should uh, they rely on the output and how have you seen them effectively um being able to assess that output, uh, base decisions on it or reject it if it doesn't um, appear to be legitimate, uh, or accurate or whatever the scenario may require. Uh, what types of steps are organizations uh using to um effectively involve LLMs? Um, given the potential risk of bad

Speaker C: um output I would say that uh the most important thing is uh to trust their internal experts. Um, I think that many companies are rushing into workforce uh reduction prematurely and exposing themselves to decisions that are suboptimal, stemming from the problem um, of uh, those who are left to make those decisions and interpret the results and structure studies to um, um being experts in other um uh jobs and not really having the benefit of decades uh of experience. So my first advice is to always rely on uh, people who are doing it in the previous waves and know exactly what it involves and what needs to be done for something to be prudent and reliable. Second, if you are in a rush and you need to just uh, form your own perception of what is reliable and whatnot, try to repeat it. That's a very simple thing that anyone can do, uh, sometimes even very inexpensively try to repeat it, uh, with some variations Try to repeat it uh, in a few days and see how repeatable that insight is. If you're just trying to do a desk research, uh, try different models from different vendors to see if they're agreeing uh, with each other or not. But if it's a uh, more involved study, um, try running it at uh, some cadence, a few waves and see if um, it's totally random or indistinguishable from uh, what the random response would be or if you really have some depth to the method that uh, that vendor or that way of gathering information entails.

Speaker B: Super awesome. So it's certainly a very interesting shift that's um, happening right now and it really appears as though it's going to only speed up. Um, there's mounting um, motivation and interest and aggressive uh adoption happening right now. Uh, following these, you know, go going through these three waves that you described and leading to who knows what they're after. And um, we really appreciate you stopping by today to give us uh, your take on what you're seeing in the industry and your expert insights. Uh, before we conclude um, could you tell us a bit more about ask your target market, uh, what you do and some of your achievements.

Speaker C: Um, thank you. The pleasure was all mine. I appreciate your questions and uh, being here um, EYTM has ah, ah had a fantastic journey and I continue enjoying every minute of it. We've been unusually uh, to other companies in our industry and in the startup world, uh self funded and uh, profitable for many years. We were honored to be named among uh Fortune 500 companies uh by Inc. 5000 or four times in different years and um, uh continue figuring out this very dynamically changing environment and industry and innovating every year. That's probably the best portion of this job to figure out what's happening and coming up with solutions that the largest CPG companies, consumer brands and financial companies in the world appreciate, trust and use every day. Thank you Thomas.

Speaker B: Great, thanks so much. And hopefully we'll have a chance to touch base on the development of uh, the enterprise consumer relationships in the future as um, more uh, agentic AI adoption has occurred and as perhaps maybe there'll be a fourth wave, you can report to us songs.

Speaker C: Looking forward to it.

Speaker B: Thank you,

Speaker A: thank you for listening. Follow Thomas on LinkedIn.