For years, transparency in CTV has been one of the industry’s most stubborn issues, with buyers frequently complaining that they don’t always know exactly where their ads are ending up. It’ll undoubtedly once again be a point of discussion at next week’s Cannes Lions. But something may be shifting on the sell-side. Srishti Gupta, Chief Product Officer at IAS, believes the participation of major publishers including Disney, NBCUniversal, Paramount, and Prime Video in IAS’s newly launched Total TV product is a meaningful signal that the industry is ready to open up.
In this interview, Gupta also discusses how IAS separates the true impact of AI from the hype, and where there’s been meaningful change in the industry since Cannes last year.
Looking at the past year since the last Cannes, how has IAS changed? What are the most significant differences in the business compared with this time last year?
Fundamentally, Cannes is a celebration of human creativity. But if you think about it in the context of the explosion of AI-driven content and the massive fragmentation of how people are consuming content today, it’s actually more important now than ever before that the creativity and the good ideas are actually reaching real people in the right environments and driving results. That is really what we’ve been focused on over the last year.
The most significant difference in how we have been approaching the industry is transforming media quality from a defensive measure protecting brands into a closed-loop performance engine. Specifically, we’ve been a lot more focused on driving value in social environments through our advanced social optimisation solutions. They provide a lot of granularity to brands, including custom brand segments designed to minimise ad waste and make sure that ads get routed to suitable feeds and environments.
The results are really compelling. We see close to a 71 percent decrease in brand suitability fail rates, an 18 percent reduction in platform CPMs, and close to a 30 percent increase in engagement. Higher media quality drives higher performance.
We’re also seeing how we can increase media ROI across the open web to enable audience-enhanced targeting, combining the depth of audience data with real-time intent of contextual signals. Ultimately, what this means for a client is an average increase of 35 percent in working media.
Looking more broadly, what in the industry do you think has changed significantly since Cannes last year? And are there any areas from last year’s Cannes where you feel we’ve not made much progress?
Where we’ve really moved forward is with respect to how important AI has become, not only in terms of experimentation, but in truly changing how we work. Real-world implementation has started to happen, largely driven by the change in consumer behaviour. Consumers have adopted frontier models and chatbots in their experience, and advertising follows where consumer attention goes. That has certainly helped accelerate it, leading to agentic workflows, automation, and much more seamless workflows across planning and activation. There has been a significant change.
As to where we haven’t seen as much change, I think we still aren’t as prepared or practical as we could be about where AI could help us with small wins. While we’re talking about fundamentally changing how we work and buy, the technology does allow us to get way more precise with that scale, and start to look for those smaller wins as well. I hope we can make more progress on that in the coming year.
AI is clearly the centrepiece of the industry right now. How is IAS separating the true impact of AI from the hype?
The way we think about it is working backwards from our clients. Does the technology ultimately connect to outcomes? Does it lead to a better result, a better return on ad spend for the advertiser? That is a pretty clear test of whether this is worth investing in right now.
We are focused on the fundamentals that don’t change. Good data has always been really important, and it is even more important in the world of AI. We want to make sure all the signals we get are at scale and that they are solid signals we can use in our models. For example, our multimodal models use 70 years’ worth of video content daily. We classify it frame by frame with 130 percent greater accuracy than other solutions. That underpinning of good data is really important.
The second thing I’m very excited about is how AI is giving us the ability to rethink classification and build it working from the brand perspective, completely tailored to brand-specific thresholds and continuously improving based on data. It is about understanding nuance and context, the way language is actually written and spoken. It’s the compounding advantage of better data, better models, and more intelligence for advertisers.
IAS launched Total TV earlier this year. For those unfamiliar, can you explain what the product is, and what issues it aims to solve?
The primary issues Total TV solves are fragmentation in the industry and inconsistency. From an advertiser’s perspective, they’re buying across dozens of streaming environments, but for each one they get an inconsistent measurement. They don’t always get transparency on a show or genre level. Compared to linear TV, they’re flying pretty blind while paying premium CPMs. That is what we solved through Total TV; we think of it as a visibility engine for the streaming era. We’ve launched with our partners in the U.S. and will be rolling out internationally in the future.
Our approach relies on direct partnerships with CTV publishers to bring linear-like transparency to CTV. We can aggregate data, normalise it, and make it consistent across show, genre, and programme levels so that a brand can easily look at their spend and see how they should optimise.
Transparency has remained a stubborn issue in CTV, and we often hear about the challenges on the sell-side when it comes to providing transparency. Total TV has an impressive lineup of TV companies involved — are sellers becoming more transparent?
I think it is a massive signal that the industry is actually ready for that change, and ready to provide that level of transparency. Total TV is supported by some very large publishers like Disney, NBCUniversal, Paramount, and Prime Video at launch in the U.S. Sellers are also leaning into this level of transparency because, ultimately, it validates the quality of the inventory they have. It builds long-term trust with the buyers that they are getting transparency. That helps them then unlock more CTV dollars, so I’m optimistic that this signal is going to get the industry leaned in even more.
What are you focused on as the Chief Product Officer at IAS?
My focus is really driven by two things. First, following the consumer, wherever consumer attention goes, and making sure we are able to provide brand suitability and performance in those environments. Second, thinking about how we accelerate our speed of innovation because we’ve been given this opportunity with AI to really reimagine how quickly we can innovate on behalf of our clients.
Following the consumer means our R&D efforts are invested where consumer attention is going. Social platforms and the creator ecosystem are growing rapidly, and as audience engagement shifts into those spaces, we are utilising frontier models to understand this complex, nuanced creator content. This helps advertisers increase their return on ad spend, protect themselves within the environment, and scale the size of those investments. Similarly, advertising within frontier models is still new, but brands are asking how they should think about brand safety and suitability in those environments.
The second area is the speed of innovation. In the world of AI, content is fundamentally changing. There is just so much content, and in order for us to outpace and understand it, we must have the velocity to process that data. As I mentioned, we process over 70 years’ worth of video content in a single day, and 280 billion digital interactions daily. Taking that massive scale and deploying AI-first workflows, such as agents, helps our clients get insights five times faster.
Those are the two areas I am focused on: understanding consumer behaviour better and making sure we have the tools to actually classify it and get insights to our clients much faster.
Looking ahead, where are the biggest challenges looming on the horizon? What topics might we be talking about at Cannes 2027?
As I said at the start, Cannes is really about human creativity and expression. In some ways, AI has made it very inexpensive and easy to produce content, but that does not mean that all of it is good. A lot of it is mass-produced, low-quality AI content, which means the great ideas will get drowned out and consumer attention is just going to be fragmented. When I think of that AI slop, it could be a massive challenge for advertisers and our industry, creating a poor environment. This is something we need to tackle head-on so that we can avoid those environments and instead find the areas where consumers are going to be leaned in and truly want to engage.
At IAS, we are already focused on this. Our low-quality generative AI avoidance solution is one example of how we are helping protect brands from low-quality content. But I think it is going to become more and more nuanced as we focus on filtering out the noise. We have seen that there is close to a 50 percent higher success rate for campaigns just by filtering out the sheer noise. Hopefully, over the next year, we can get much more nuanced around this and really be able to separate the noise from what is true, highlighting the great ideas and great creative.
Follow VideoWeek on LinkedIn.
