All Things Insights sat down with Chavda as he explained how agentic workflows can automate complex research processes like innovation concept development, connecting multiple AI modules to create dynamic systems that complete tasks in hours rather than weeks.
The conversation covers the shift in industry attitude from skepticism to adoption, addressing concerns about job displacement while emphasizing the need for researchers to develop new skills. He notes that while AI accelerates research processes dramatically, the need for insights and primary data will increase, not decrease, making speed and business application more critical than ever.
The Next Wave of AI
All Things Insights: We’re here on the second day of TMRE 2025, and we have the pleasure of being with Yogesh Chavda, consultant and educator, Founder of Y2S Consulting. Thanks so much for coming on and talking with us a little bit about insights today. You had a session on an important topic right now, agentic AI. Can you tell us a little bit about this next wave of AI taking shape?
Yogesh Chavda: Thank you for having me. Agentic models or workflows basically are processes that you can actually automate and they are dynamic automation systems. You could have specific GPT modules that do very specific tasks, but then you can connect them together to give you a workflow.
To give you an example, on the client side, brand teams work on creating innovation concepts, for example. And it’s a process. You have to understand the market. You have to identify an idea that’s in a white space opportunity. You then have to start drafting concepts. Maybe you’ll do some focus groups. You have to get feedback from your boss. Probably rewrite the concepts and that whole process takes like weeks or months in some cases.
But now you can actually build an agentic model that does all those steps for you. Step by step, the way I just described it, and all of a sudden you now have a workflow that’s completely dynamic in nature. All you’re doing is putting in your inputs. The model will basically do all those steps for you, do all the analysis, the writing of the concepts, get feedback, and then rewriting the concepts. That’s an example of an agentic workflow.
Connecting it All Together
All Things Insights: I’ve also heard agentic AI viewed as an additional assistant in a way as well, kind of like your own personal agents that are helping the process along.
Yogesh Chavda: Yes, but there’s a difference. So you have custom GPTs, which are very specific agents that do very specific tasks. You could have an agent that says, help me craft a presentation, for example. And all it does is do the quick presentation. But it’s a single activity system, if you want to call it that.
In an agentic world, what you’re doing is you’re saying, OK, I’m going to create the content, but I’m going to now put it onto PowerPoint. And then I want to have images in there and I want it to basically be contextualized in a particular way and I want it to be emailed back to me. Those are all steps. You can’t do that in just one module. You have to connect different modules together. For example, if I want to drop it into PowerPoint, that’s a separate module that is not even ChatGPT. That’s a totally different tool. So to connect those things together, that’s where you have to have other platforms that connect these things together. Now I will say that both Gemini and OpenAI have launched something on their platforms where you can do this too. It’s relatively new. Most people don’t even know that it exists on the platform right now.
Building Your Skill Set
All Things Insights: Let’s take a little bit of that view of this from 20,000-feet up. There’s a lot of that fear factor in the community. I know you have a technological stance, a viewpoint to embrace these changes. What can the researcher do in their day to day as far as experimentation? Where do you feel the community is at this moment in time?
Yogesh Chavda: I certainly sense a shift from a year ago. Last year, I felt there was a lot of hesitancy and skepticism about a lot of different things. Whether it was about AI or synthetic data or even custom GPT agents, people were looking at it and saying, really do I need this, or will it actually take away my job and all that kind of stuff.
This year, what I’m seeing is that there’s a lot more companies and people embracing it, which is a good sign. They may not be completely there to the agentic world just yet, but there is a shift that I’m noticing. Is the fear factor still there? Yes. Just look at the announcements that have come out. Amazon announced 30,000 layoffs. Paramount announced a couple of thousand people as well.
It’s becoming a real thing. And my point of view in all of this is that, yes, there will be jobs that’ll be disappearing, but at the same time, there will be new jobs that are going to be emerging. And they will require a very high level of skills in knowing how to do these kinds of things that I’m talking about right now.
Fom an industry and a community perspective, I would be encouraging everyone to aim up and don’t wait. Don’t sit on the fence. Pick a swim lane and start doing something and start becoming an expert in something because you will find that that will eventually help you in the long run. Will it save your job today? I don’t know. Because there are decision makers looking at the situation and saying, if we can get cost savings or if we can get a margin improvement by using AI, senior leaders are going to do it. And there’s nothing that you can do in that situation.
But what you can do is build your skill sets up while you have a job or even if you don’t have a job, start building your skill sets now so that you can start influencing companies to say, you will need these kinds of skills in your company. I’ve got it.
Accelerating & Activating Insights
All Things Insights: Speaking of that new skill set for today’s modern researcher, someone mentioned kind of being the Swiss Army knife of research and just having all of these new skills. Is it driven more by business strategy now? Is that the key goal at the end of the process, to transform insights into business strategy?
Yogesh Chavda: It should always be that, actually. I would say regardless of AI, that has always been the rallying cry. Quite frankly, what I have found within the industry is that a lot of people still focus on the methodologies as opposed to focusing on the business. The more you try to translate what your unique skill set is and, from an insights perspective and apply it to the business, that’s when you show value to the company and you become invaluable to the company in a way because they see that you’re driving revenue or you’re driving profits or both. When they see you as a cost, then it’s easy to cut.
AI is just basically accelerating and amplifying that issue for everyone right now.
All Things Insights: This dovetails into your other session, which had to do with technology and AI as well. As you phrased it in the session, it was about design, validate, and activate. There was a mantra of churning insights into business strategy as well. Just share a little bit more about that session?
Yogesh Chavda: Quick backstory. The thinking behind what I shared yesterday actually started back in 2017 when I was working at Spotify. I designed a framework at Spotify that I have now basically recreated in the context of a custom GPT agent. In essence, what I’m doing is if a brand wants to create personalized messaging, you have to be able to understand what are those moments in the consumer’s life when they would be most receptive to hearing the message? And it’ll vary depending on the time of day or the activity you’re doing with the your frame of mind that you have. My GPT agent actually helps to determine all that, identify those kinds of moments, and then it determines what kind of messaging would resonate for your brand in the context of that moment.
That’s part one. But the second part that I’ve also built into that is that if I want to activate that message across platforms like Meta or Google or TikTok, I’ve uploaded all the documentation for those platforms so that it’ll tell you exactly that, OK, if you want to activate on TikTok because that moment is relatively important to you from a TikTok perspective, it’ll tell you exactly how to go activate it. It connects you from designing to validating to activating.
Picking Up the Pace
All Things Insights: That taps into the whole speed of business right now. Are you saying you can do something like that in mere days or hours?
Yogesh Chavda: Hours. Literally, it’s hours. Speed has definitely increased. If you think back to the 1990’s, for example, when we shifted from paper surveys, which I remember doing when I was a new hire to online surveys. We were like, oh my gosh. This is now really accelerated. We were all rejoicing in that. Well, some of us were. Now it’s not weeks or days anymore. It’s now literally hours. The pace has certainly picked up.
But I want to say one more thing. The need for insights and the need for primary data is actually going to increase, not decrease, because of AI. And people miss that point very often actually.
All Things Insights: The speed of insights; it’s the real-time nature of the business.
Yogesh Chavda: That’s exactly right.
Editor’s Note: The next TMRE will take place on October 5-7, 2026, and will be co-located with Content Marketing World at the Colorado Convention Center, Denver, CO.
