CTV Gets Smarter: James Grant of Equativ on AI, Identity & Outcomes

Equativ's James Grant on CTV: What's Real, What's Marketing

Equativ SVP James Grant sat down for a sponsored conversation on CTV monetization, covering live-sports streaming, pause ads, and the company's AI planning tool, Maestro. Strip the sponsorship layer and two things are worth your time: a live-sports war story with a concrete lesson, and a quiet admission that identity outside the walled gardens is still broken.

The AI pitch is early. Grant bills Maestro as an AI media planning agent, then walks it back to "workflow efficiency, starting to expand into insights." That's a suggestion engine trained on Equativ's own campaign data. Every SSP with a data warehouse will ship one this year. Treat any AI planner pitch in 2026 as workflow software until the vendor shows you incremental lift on your inventory, not a demo.

The live-sports story is the useful part. Grant describes a U.S. broadcaster that ran blank slates during the World Cup because it sold inventory on linear-style upfront share-of-voice deals and had no real-time programmatic fill when the live audience overshot the forecast. Sales teams priced live sports like linear TV because that's how buyers negotiate, and when the audience beat forecast, there was nothing to run. Grant's practical lesson: live-sports streaming requires a 12-to-18-month infrastructure commitment before the event, not a last-minute bolt-on. Anyone signing live rights now and skipping that build is setting up the same failure.

Pause ads and home-screen formats are the near-term opportunity. You add sellable inventory without adding mid-roll interruptions, so you're not diluting CPM on existing supply. OS-level partnerships like Titan are moving these formats toward standardized auctioned inventory, which means SSPs and DSPs that build format support early get first access. Sharethrough's attention research claims a 50%-plus lift from well-executed creative in pause placements, which is at least grounded in years of prior work.

CTV performance claims need scrutiny. Grant argues that CTV can now confirm a conversion, and buyer demand for that signal is real. But a visible conversion and a conversion CTV caused are different things. Attribution and incrementality are not the same, and Grant blurs them. Ask for the incrementality test before you believe the outcome story.

Identity remains the actual moat. The through-line across the whole conversation, and the least sponsored thing Grant said, is that deterministic identity outside the walled gardens is still broken. In GDPR markets like Germany, the ability to stitch consented first-party data across screens is worth more than any planning tool Equativ or anyone else ships this year. That work doesn't get the headline, but it's where the durable advantage sits.

Our call: During at least one major live-sports streaming event in the 2026-27 window, including the 2026 FIFA World Cup final rounds or the 2027 Super Bowl streaming window, a U.S. or European broadcaster will again run blank slates or house ads because it sold inventory on linear upfront logic and lacked real-time programmatic fill for an audience that overshot forecast. The commercial model that caused the World Cup failure hasn't changed in one cycle, and the infrastructure fix takes 12 to 18 months minimum. We revisit by February 2027.

Your draft

A sponsored conversation with James Grant, SVP at Equativ, on where CTV money and effort should go: linear-to-streaming migration, performance TV, and AI media planning. Strip the sponsorship gloss and the useful signal is a live-sports war story and a candid admission that the identity problem everyone stopped talking about is still broken.

This is a briefing, not a decision. So the frame is: what in this conversation should a publisher, SSP, DSP, or agency operator actually act on, and what should they ignore as vendor marketing? Reversibility is high for the reader. Nobody is being forced to buy anything. The forcing function is the next live-sports rights cycle and the next round of "AI planning tool" pitches landing in your inbox.

The Skeptic. The headline says "AI agents." Grant himself walks it back to "workflow efficiency, starting to expand into insights." That gap is the whole story on Maestro. A suggestion engine trained on your own platform's historical campaign data is a fancy autocomplete for media plans, and every SSP with a data warehouse will ship one this year. The "40 to 45 percent of CTV content with program signals" number is Equativ's own inventory, sold by the man who benefits from you believing contextual targeting is ready at scale. For an informed outsider: the vendor is selling a search box and calling it a robot planner.

The Operator. The MagentaTV World Cup case is the part worth your Tuesday morning. Grant's real lesson is the 12 to 18 month lead time, and that project management beats the tech stack. That is a warning to anyone signing live-sports rights: your SSAI and real-time fill have to be committed a year out, not bolted on. The failure mode he names is concrete. A U.S. broadcaster ran blank slates during the World Cup because it sold inventory on linear-style upfront share-of-voice deals and had no real-time programmatic fill when the live audience overshot the forecast. That is a revenue hole you can drive a truck through, and it comes from the commercial model. For an informed outsider: they sold the ad slots in advance like TV, the crowd showed up bigger than expected, and they had nothing to run.

The Customer / End User (the buyer and the viewer). Two different customers, two different reads. The buyer wants outcomes, and Grant's claim that the biggest screen can now confirm a conversion is the real pull in CTV. That's genuine demand, not projection. The viewer read is pause ads. Grant is right that an ad served when the viewer chose to pause is less hostile than a mid-roll interruption, and Sharethrough's attention work claiming a plus-50% lift from creative like a sports ticker is at least grounded in a decade of research. For an informed outsider: the ad shows up when you hit pause, so it annoys you less, and that makes the inventory easier to sell without wrecking the experience.

The CFO. Pause ads and home-screen formats are the cheap growth here. You add sellable inventory without diluting your mid-roll CPMs, because you are not stuffing more interruptions into the show. Contrast that with live sports, which is a capital and calendar commitment measured in years and only pays if you win rights and build the fill infrastructure to match. The AI planning tool barely moves the P&L yet. If it saves planner hours, fine, but nobody is retiring a headcount line on a suggestion engine in 2026. Spend the attention on the format expansion and the identity work. The demo can wait.

The tensions

Where do thoughtful operators split?

First, is CTV performance real or is it a measurement story dressed for the buyer. The Operator and the buyer-Customer say the conversion signal is the unlock. The Skeptic says every SSP claims outcomes now, and a conversion you can see is not the same as a conversion CTV caused. Attribution and incrementality are not the same thing, and Grant conveniently blurs them.

Second, spend the year on live sports or the quarter on pause ads. The CFO wants the low-friction format win. The Operator sees live sports as the moat, but only for platforms that can eat an 18-month build and a rights bill. Most readers are not Deutsche Telekom.

Third, and the one Grant surfaces almost as an aside: everyone chases AI and outcomes while deterministic identity outside the walled gardens stays broken. That through-line runs the length of the episode, and it's the least sponsored thing he said.

Synthesis

This hinges on separating what Equativ is selling from what the conversation reveals. The sponsored layer is the AI planning agent, and it is early. Treat any SSP or DSP "AI planner" pitch this year as workflow software until it proves predictive optimization on your data, and make the vendor show you incremental lift, not a chat box.

The unsponsored layer is more useful. Live-sports streaming exposes a commercial-model failure, not a technology failure. Broadcasters selling live inventory on linear upfront logic will keep running slates when audiences overshoot, and that is a share shift toward whoever can fill programmatically in real time. Pause ads and home-screen supply are moving toward standardized auctioned inventory via OS partnerships like Titan, which means SSPs and DSPs that build format support early get first dibs. And the identity layer remains the real moat. In GDPR markets like Germany, the ability to stitch consented first-party data across screens is worth more than any planning bot.

De-risk it this way: before you buy an AI planning tool, ask for the incrementality test. Before you sign live rights, commit the fill infra a year out and stress-test the commercial model against audience overperformance, not just under-delivery. Before you believe the contextual-signal numbers, ask whether they describe your inventory or the vendor's.

The prediction

The most gradeable claim in this episode is the live-sports fill failure, because Grant named it happening already and the mechanism is structural. Prediction: During at least one major live-sports streaming event in the 2026-27 season, including the lead-up to and broadcast of the 2026 FIFA World Cup final rounds or the 2027 Super Bowl streaming window, a U.S. or European broadcaster will again run blank slates or house ads because it sold inventory on linear-style upfront deals and lacked real-time programmatic fill for an audience that overshot forecast.

Confidence: Medium. The commercial model that caused the World Cup fill failure hasn't changed in one cycle, and the infrastructure fix takes 12-18 months minimum.

Why: A U.S. broadcaster ran slates during the North American World Cup because it priced live inventory as linear-style upfront share-of-voice and had no real-time fill when the live audience beat forecast. That is a commercial-model problem, and commercial models move slower than one rights cycle: sales teams still price live sports like linear TV because that is how buyers negotiate and how guarantees are written. Building programmatic fill for live at scale is a 12-to-18-month infrastructure job, so most broadcasters heading into the next major tent-pole window won't have closed that gap. The opposite outcome, clean sold-out real-time fill across every major event, would require the entire broadcast sell side to have re-architected both pricing and infrastructure inside a single year, and no part of the industry has that track record.

Revisit by 2027-02-28: We're right if a named broadcaster runs slates, house ads, or visible unfilled inventory during a major live-sports streaming event through the 2027 Super Bowl streaming window, reported in trade press or flagged by viewers. We're wrong if no such fill failure surfaces at any major live-sports streaming event in that window.