← The Digital Executive
The Digital Executive
Vivek Kumar on Reimagining Market Research Through AI-Powered Audience Simulations | Ep 1064
00:10:04
Loading summary
Transcript14 lines
- [00:00]
A
Foreign.
- [00:09]
B
Welcome to Coruscant Technologies, home of the Digital Executive Podcast. Welcome to the Digital Executive. Today's guest is Vivek Kumar. Vivek Kumar is the CEO and co founder of SocialTrade, the world's first audience behavior simulation engine. With over 15 years of experience in business strategy and and AI driven automation, Vivec has helped scale products and drive market expansion across the public, financial and tech sectors. He began his career in the UK public service at HM Revenue and Customs and later transitioned into investment Banking at the Royal bank of Scotland where he also earned his Master's in Finance. From there he went on to co found a sustainable waste management startup and was recognized in the Shell Livewire Young Entrepreneur Award. After several ventures, Vivek relocated to Australia and co founded Social Trait alongside Siraj and Gopal Kumar. In 2022, the company was named one of the 20 most innovative startups in the industry by Berkeley Skydeck. Well, good afternoon Vivek. Welcome to the show.
- [01:08]
A
Thank you.
- [01:09]
B
Absolutely. I appreciate you making the time my friend. You are currently in Melbourne, Australia. I'm in Kansas City, Central of the United States. So we've got quite a bit of a time zone difference. So again I appreciate you making the time at that time of the evening very early here in Kansas City. So Vivek, let's jump into your first question here. You've had a fascinating journey from public service to AI innovation. What inspired the leap from HMRC and investment banking to founding a research tech startup like SocialTrade?
- [01:43]
A
Picture this Brian. Four years ago SocialTree's co founder and CTO Suraj posted something on social media and that didn't go well and with a certain group of audiences. And within hours he was buried under thousands of angry comments. And that single moment lit the fuse for social trade. I had spent years in data heavy roles as you mentioned, in the HMRC tax office and then investment banking. Always trying to make numbers speak human. But Surat's backlash showed me a bigger gap. Brands have no safe way to see real reactions before they gamble millions on a launch. And marketing teams burn weeks recruiting focus groups only to watch campaigns flop in the world. I moved to Australia. I was introduced to Suraj the AI wizard and Gopal, the operational A's. Over one late night coffee, we asked what if we could build a virtual world populated by AI consumers where you can preview emotion, outrage or delight before you hit publish. That idea became Social Trait. Thousands of AI agents who read, watch, share and argue like real people, letting brands rehearse ideas the way Pilots use Flight simulator and today our clients swap six week research cycles for 48 art simulations, cutting risk money and sleepless nights. So that the leap wasn't about leaving one career for another, it was about chasing a bigger purpose. Showing companies what people will feel tomorrow, today.
- [03:19]
B
That's amazing. Thank you. Yes, of course. We always talk about the story, the genesis of launching a platform, a product or a solution to a problem. There's some sort of story behind it. Those are always fun and exciting and interesting most of the time. And yours is the same. It's just amazing. I love the story of how we kind of learn the hard way, right? We post something and then of course we get some backlash. But I love your product, obviously allowing AI agents to do some simulations prior to sending out a post. And I think that's important. That's kind of the safe way. But also get better engagement as well. Vivek Social Trait blends generative AI with community simulation can you explain how your platform works and what makes it different from traditional market research tools?
- [04:03]
A
We have built like AI agents. You're hearing about AI agents everywhere, right? And people talk about one AI agent. So a few AI agents, but we have built over half a million AI agents in the virtual world right now. So for our clients we build AI agents that behave like real people. So the goal is to allow business to create virtual communities consisting of thousands of AI agents that mimic their target audience. So they have demographics, psychographics and they have salient history just like us. They can watch, listen, read any kind of digital content and consume information, you know, just like you and me. And then we group those agents into communities that mirrors brands, actual segments like loyalty moms in Minneapolis, sneakerheads in la, et cetera. Teams can run experiments like large scale surveys, simulated surveys, focus group image or video rackings, attention heat maps, just like any attention heat where you can actually literally see where your target audience are looking at an image or a video and inside this synthetic world and watch opinions form in real time and the output is deeper, wise actionable insights at a fraction of a cost in time. So if you think of for example like a global food brand, cut a 30 day position study to few days choosing the winnings messages after 1,500 AI shoppers ranked 13 taglines. So the question is why? It's game changing. 48 hours cycles versus 6 weeks panels reusable virtual audiences, so no recruitment fees, so there is no research fatigue. And you can uncover deeper whys and action insights which you cannot do with the traditional Methods.
- [05:47]
B
Awesome. Thank you, I appreciate that. Building a half a million agents that behave like humans in this synthetic world, as you call it. These agents obviously have similar backgrounds to humans, interests, likes, et cetera. But what I really like, and I think it's going to be attractive to your customers is, you know, no more surveys, no more multiple week surveys, no focus groups, no research fatigue. Of course you're going to save a lot of money on that. So I appreciate you breaking that down for us and for our audience. Vivek, many startups struggle to scale AI solutions. What challenges did you face in developing social traits, virtual focus groups, and how did you overcome them?
- [06:24]
A
Obviously there are a lot of tech related issues, but it was something different. To be honest, our biggest challenge wasn't technology. Yes, they were getting the right GPUs, getting all those things. Infra ready is a common among most of a tech startup at the moment, but it was actually keeping up with the demand of customers, you know. And when the ChatGPT went mainstream, our inbox lit up with Fortune 500 leaders who who all wanted AI pilot. Yesterday, bandwidth became a bottleneck. We had world class engineers, but not enough market research muscle to run 10 enterprise studies at once. So we hired seasoned insight leaders from top brands to not only kind of help us to execute some of those early pilots, but also to give us feedback on our products so we can scale quickly. They translated client briefs into rigorous study designs while our engineers kept scaling the agent platform. We also built self serve dashboards so clients could experience platform firsthand without waiting on our team result. In six months, we went from handling three concurrent pilots to 20, did more than 300 research projects while keeping the turnaround under 72 hours. And a key takeaway in B2B AI, you scale trust first, software second.
- [07:41]
B
Now that's amazing. B2B, you scale trust first. I love that and not the software. I think that's important and I think that's going to be foundationally throughout anything we do going forward, especially in this disruptive market we have with AI. It's interesting that I've heard bottlenecks, but your biggest challenge was the bottleneck keeping up with the demand. It wasn't the technology. A lot of time, technology can be the issue. I appreciate you sharing some of the struggles and challenges you had early on. Vivek, the last question I have for you this evening, your time. What's next for Social Trait? Are there any upcoming features, markets or partnerships that you're particularly excited about?
- [08:20]
A
I think there are plenty of them, but yeah, we're building a way for brands to a b test the virality itself. That is one of the biggest challenges in the industry. To understand the spread, we are launching a social media simulation world populated by millions of AI agents who scroll like comment and create content 24 bars old. Just like us. They consume the news information and they just behave like that. But the world with only AI Personas in it. So the usage is a marketer can drop a post or an ad into this universe and watch in real time how fast it spreads, who boosts it, where backlash sparks. Because the agents behave organically, we can forecast reach curves, influencer lift and crisis flashpoints before a single real consumer sees the campaign. On the partnership front, we are integrating this engine with top tier retail media network so creative can be fine tuned moments before it goes live. Our North Star is an audience behavior simulation layer that every marketer taps before spending a dollar in the wild.
- [09:26]
B
That's awesome. I appreciate that and you're taking the AB testing to the next level. But I like that you're going to create this social media virtual world with all these AI personalities and we're going to get some amazing insights through this testing before it's even launched out to the real world. So I appreciate that and I'm excited to follow you going forward and looking forward to have you back on the podcast. I appreciate your time today, it was such a pleasure and I look forward to speaking with you real soon.
- [09:52]
A
Yeah, same here. Thank you so much Brian. So nice talking to you.
- [09:56]
B
Bye for now.