Programmatic DSP Viant just closed on its $40 million cash-and-stock purchase of attention measurement provider TVision Insights.
As an attention measurement provider, TVision delivers second-by-second, eyes-on-screen attention, co-viewership and in-room presence signals for TV, offering independent measurement.
The acquisition helps bolster the foundation for Viant’s larger AI-powered platform and vision for fully autonomous advertising, including the so-called intelligence layer strategy underpinning the platform where data signals are used to optimize campaigns in real-time.
The TVision purchase also serves as a more direct challenge to walled gardens like Google and Meta, which Viant claims suffer from grading their own homework with self-measurement.
Viant - which generated record quarterly connected TV ad spend in Q4, with the category accounting for 46% of the DSP's total ad spend in the period - has already been working to attract advertisers and performance marketers that usually spend in walled gardens by following a similar playbook for open internet and CTV advertising in terms of providing brands with conversational agentic AI tools for campaign planning, optimizations and outcomes.
And with TVision data and signals integrated directly into the Viant DSP, it’s also positioning the purchase as a shift away from impression-based metrics toward attention-adjusted outcomes, including with a new “attention-adjusted CPM.”
StreamTV Insider spoke with Viant CMO Jon Schulz ahead of and since the TVision acquisition announcement - read on to see what he had to say about what the company is seeing in terms of marketers’ readiness for agentic AI tools, the DSP’s challenge to walled gardens and how TVision comes into play to support larger company goals.
Marketers cautiously ready for agentic AI, but prefer controls
In an interview with StreamTV Insider ahead of the TVision buy acquisition, Schulz described what we’ll call a “cautious readiness” among marketers when it comes to using agentic and other AI tools. But he suggested they aren’t shifting into autopilot mode and instead prefer to keep hands on a steering wheel for now.
On the path to Viant’s vision for autonomous advertising, latest and last pillar in the Viant AI strategy is what the company dubs AI Decisioning with a product called Outcomes, powered by the company’s AI Lattice Brain and intelligence layer.
It’s a fully autonomous AI decisioning product meant to plan, build and execute campaigns that are designed to produce the optimal brand result (or outcome) based on a few simple inputs and prompts – like goal, CPA, budget, and so on - by advertisers.
Schulz likened it to Viant’s version of Performance Max or Advantage+ but for the open internet.
The Outcomes launch in January followed the initial introduction of Viant AI in 2024, which brought LLM-based AI agents ChatGPT-style into the mix for optimized planning and budget allocation recommendations, distilling trillions of possible combinations across the open internet via simple inputs and conversational prompts
There’s been a lot of buzz about both agentic AI and “outcomes” as of late with multiple companies introducing their own tools and capabilities. But the technology is new and still emerging, so perhaps understandably it isn’t an instant overhaul of long-standing processes – particularly when it comes to levels of trust and confidence needed by advertisers to allow AI to make actual ad buy decisions.
Marketers do appear keen to make the most out of budgets and simplify the fragmented and complex digital ad space, per Schulz, but they also know their brands well and want to maintain a level of control and guidance over still evolving AI tools.
When talking to customers, he noted that some are very interested and find the AI tools impressive but also acknowledged that fully handing the reins over to AI for things like ad buy decisioning (as opposed to planning or recommendations) isn’t yet something all feel comfortable with - – a sentiment the CMO understands.
“I don’t know if I’d step into a Waymo self-driving cab,” Schulz said, offering an example of a separate autonomous technology. “I want to have my hands on the steering wheel and my foot on the pedal. I just don’t have that level of trust yet.”
Separate survey findings from MediaOcean support the notion that marketers are interested in CTV (and plan to increase spend on the channel) and AI, but challenges around things like fragmentation and measurement remain. And marketer respondents signaled a need for trust and a level of control when it comes to AI tools.
Per the January findings, 49% cited complexity of cross-channel measurement and optimization as their top concern for media and marketing initiatives. And 43% reported concern over balancing AI adoption with brand safety, accuracy and creative control, while 42% worry about utilizing AI-driven automation without losing human oversight or quality.
However, for Viant, TVision represents another step in building that trust, where it also integrates with other data feeding into the advertising platform and intelligence layer that underpins the AI systems.
That includes integration with content and contextual signals strengthened by its earlier purchase of IRIS.TV and its IRIS_ID, as well as identity resolution with Household ID audience signals. Together with AI, this is meant to drive smarter ad targeting and measurement for advertisers on the DSP.
And in terms of scaling technologies, Schulz told us that when IRIS was first acquired in 2024 its content ID was available in less than 10% of the bidstream (or among bid requests, less than one out of 10 had an IRIS_ID attached) – now that’s grown to 50% or one out of two bid requests, and Viant sees room to scale further.
How TVision buy bolsters Viant’s AI platform
With marketers’ cautious readiness around AI in mind, StreamTV Insider asked Schulz how the TVision purchase will make an impact.
Per the CMO, Viant’s seeing “strong interest in AI-powered planning and optimization, but most advertisers still want a level of control.” What TVision does “is strengthen the foundation underneath that AI.”
“AI is only as effective as the quality of the data it learns from. By adding second-by-second, person-level attention signals into our Viant Intelligence Layer, we are giving advertisers more transparency into what is actually working,” Schulz told us. “That helps build trust, not by removing the steering wheel, but by making the system smarter and more explainable.”
And in his view, over time, that type of independent, verifiable signal is what will speed up advertiser comfort with more autonomous ad buying.
“It is less about forcing automation and more about proving performance in a way that advertisers can validate,” he said.
Validating performance may be all the more important as Viant’s AI platform, and Outcomes specifically, opens the door to a new pool of potential performance-focused advertisers and ad spend for the DSP.
Viant’s performance marketer push, expanded TAM
Viant isn’t the only one bringing agentic AI advertising solutions to market, but its earlier Viant AI tools and the Outcomes product also helps expand the company’s total addressable market (TAM) to include performance-based marketers – aka those more directly or typically seeking sales as a result of ad spend (as opposed to more upper-funnel aims like awareness and brand building).
As we’ve covered before, it’s a potentially large and lucrative market as these advertisers often include DTC, e-commerce and smaller-to-mid-sized brands that usually spend within walled gardens like Google and Meta.
Those walled gardens have introduced their own self-serve and agentic AI tools (like Google’s Performance Max and Meta’s Advantage+).
Schulz gave credit to Google and Meta, which have had success with their respective products, in terms of training the market to have more confidence and trust in that “if you want to deliver performance and results, this is a very efficient way to do that.”
But CTV has potential and efforts by multiple players, including Viant, are underway to make the channel more of a full-funnel and performance marketing vehicle and where advertisers can find new customers for their wares.
That doesn’t necessarily mean delivering transactions or direct sales through CTV ads (although the capabilities are there with the rise of interactive and transactional or shoppable formats), as Schulz believes TV is still largely a mid to upper-funnel marketing vehicle, but bringing performance marketers into the fold to achieve their goals through the powerful video medium on the home’s largest screen.
“To grow the business, you need new customers, and so that's really where CTV comes in to the mix [for performance marketers],” Schulz said. “It is the driver of that demand generation and making people aware. You learn about new products more through the sight, sound and motion of video and the advertising that you see there.”
And Viant saw an opportunity via Outcomes to use a similar playbook outside of walled gardens for the open internet, where it can offer brands access to inventory from major CTV publishers like Disney, Peacock, Paramount and so on.
Similar to the conversational-back-and-forth of the ViantAI planning tool, with Outcomes marketers can enter parameters, like goal, CPM, budget and then the Viant Lattice Brain optimizes towards that goal.
“Those are the applications that really are the difference maker when you go beyond what a human would even be capable of doing,” Schulz said regarding AI.
Even if there was desire, he noted that a human simply couldn’t sit there and figure out thousands of optimizations a day when it comes to bidding, publisher conversions and combinations of other factors to yield the best possible campaign performance or outcome based on the goal and parameters.
The company sees a lot of potential and upside with Outcomes, but for now it’s aiming to meet customers where they are. Most of whom are ;still “hands on keyboard” or using self-service tools and managing optimizations themselves in familiar and comfortable fashion.
And with TVision Viant’s making a more direct play and challenge against those walled gardens, which the company claims suffer from the practice of grading their own homework when it comes to measurement.
TVision buy helps Viant challenge walled gardens
In closing the TVision buy, the DSP says it can deliver independent, market-wide measurement across linear, CTV and walled gardens that’s free from any single platform’s self-attribution bias. In turn, that means it can direct advertisers to inventory that drives optimal outcomes, the company contends.
Specifically, with the acquisition completed, Viant’s platform will be able to deliver the following:
- Attention-adjusted CPM: a first-of-its-kind metric that prices CTV and linear inventory based on verified in-room presence, co-viewership, and eyes-on-screen attention rather than served impressions.
- Market-wide measurement: independent attention signals across linear, CTV, and walled gardens, free from any single platform’s self-attribution bias.
- Integrated intelligence: attention, co-viewership, and in-room presence signals will flow natively into Viant’s buying platform, alongside Household ID audience signals and IRIS_ID contextual signals.
“Today marks the moment advertisers don’t have to accept self-measurement as the TV industry standard,” said Tim Vanderhook, CEO and Co-Founder of Viant, in a statement. “TVision completes a thesis we’ve held for a long time: that attention is the only currency that actually matters in television, and we can now act on it in real time inside the buy.”
StreamTV Insider also asked about how TVision impacts Viant’s ability to cater to performance marketers that typically spend in walled gardens and attract their budgets.
According to Schulz, the purchase and integration will directly inform attention adjusted media buying and optimization.
“This will impact full-funnel TV buying in a meaningful way as clients are focusing on ad impact and outcomes over simply delivering impressions,” he said. “This is key for both brand building and performance as an unseen ad does not impact either. Like we have seen with historical innovation in ad tech, the investment will migrate to where there is proven impact. Attention is the new currency.”
Reframing how ad inventory is valued with “attention-adjusted CPM”
TVision is also changing how Viant wants ad inventory to be valued.
As Schulz described, instead of pricing media only on delivered impressions or “perceived application-level value”, Viant can incorporate typically tough to gauge signals like in-room presence, co-viewing and eyes-on-screen attention to better understand the true value of an impression.
“What that means in practice is that some inventory that looks more expensive on a traditional CPM basis may actually be more efficient when you account for how many people are truly engaged,” he explained. “Conversely, inventory that appears cheaper may be less valuable if attention is low.”
For advertisers, Viant believes this creates a clearer path to optimizing for outcomes by aligning ad spend with attention and performance rather than just reach.
“Over time, this will lead to more efficient pricing dynamics across the ecosystem, where high-attention inventory is appropriately valued and budgets flow accordingly,” Schulz said.
TVision implications for co-viewing, live sports streaming
Another aspect about CTV advertising is that two factors – co-viewing (ie: more than one person in the room watching) and attention (ie: the TV is on but are people actually watching) historically have been hard to measure.
This is where Schulz described TVision as “truly unique and differentiated” thanks to its nationally representative panel that captures co-viewing and attention at a person-level second-by-second.
“That allows us to move beyond platform self-measurement and proxy metrics to understand how many people are actually watching, how engaged they are, and how that varies across content and ads, including high-attention environments like live sports,” he noted. “For advertisers, that translates into more accurate valuation of inventory and better decisioning around where to allocate spend.”
There are also implications for advertisers around live sports – where premium inventory continues to migrate to CTV, such as this year’s Super Bowl and Winter Olympics streaming on NBCU’s Peacock, for example.
Live sports usually mean high advertiser interest but historically there has not been a ton of visibility into factors like co-viewing and attention.
“In live sports specifically, where co-viewing is high and attention can fluctuate in real time, these signals create an opportunity to make smarter buying decisions based on actual engagement rather than assumptions,” Schulz said.
So while it may not be today, Viant has confidence that the future is autonomous advertising and does expect marketers to get comfortable with agentic AI tools eventually.
And feels it’s far ahead of the game when they do.
“We just want to be ready for the next phase and we want to be thinking about that well in advance of that moment arriving,” Schulz said.