Fresh off Innovid launching NIVO AI globally, B&T sat down with Women Leading Tech winner and Innovid’s APAC general manager, Georgia Brammer to discuss the adtech platform’s hottest technology.
B&T: Innovid launched NIVO AI globally this week. In plain terms, what is it?
Georgia Brammer: NIVO AI is Innovid’s AI intelligence layer. It does two jobs. It powers a set of agents that handle the actual work of running a campaign across creative, delivery, measurement and optimisation, and it runs the orchestration that connects those agents so the whole thing moves as one. The short version: it turns a brief into live, optimised campaigns in a fraction of the time it takes today, with people still making the calls that matter.
B&T: Australian teams tend to run lean, whether that’s a small in-house team wearing every hat or activation stretched across offshore and global ops. What does NIVO actually change day to day?
GB: In Australia local teams are smaller, so the people doing the strategy are often the same people doing the manual setup, trafficking and QA. Time and time again we hear the same pain from teams working with offshore or global ops: time delays and human error, both coming from the same place, activation pulled away from the people who own the decisioning and strategy.
With the Campaign Trafficking Agent, a local marketer can upload an approved plan from their director or CMO, an email, a media plan, a spreadsheet, and NIVO builds it into complete campaigns in the ad server in minutes, ready to check.
In one live test, a large global retailer cut setup by more than half, and a trafficking and QA workflow that ran an hour and forty minutes came down to forty-five. That time goes back to the work that actually needs a person who understands the content, the context and what the client is trying to do.
B&T: A lot of what’s been sold as AI in advertising so far has been recommendations and dashboards, not execution. What’s actually been missing?
GB: Execution. Most of it is reflective: it tells you what happened, or recommends what to do next, and then a person still has to go and do it across a dozen disconnected tools. Creative and gen-AI use cases are everywhere, but the question I’d ask is what’s actually changed the infrastructure, not just run as another test on a campaign.
Our 2026 Advertising Outlook reports fragmentation across platforms and publishers is marketers’ single biggest concern, named by 56 per cent. They’re happily using AI for research and analysis, but only 19 per cent have it running campaign orchestration, and just 10 per cent say their adtech is genuinely unified.
What’s been missing is the infrastructure that connects intelligence to action and does the work inside the platforms advertisers already run on. There’s a data side to it too. Outputs are only as good as the inputs, and one extra space, one spelling error, one rogue creative name can pull a whole measurement taxonomy apart and you lose the data that mattered. Agents only earn their keep if the plumbing underneath them is clean.
B&T: With the big platforms pushing their own AI to keep spend inside their walls, and first-party data rules tightening, why does an independent layer matter more now?
GB: Because the walls are getting higher. The big platforms are pushing their own AI to keep spend and decisioning inside their walled gardens, which is fine right up until you want an honest read on what’s working across all of it. The lines between the buy side and the sell side are blurring too, agency models, the role of the DSP, agentic media buying, all shifting, with more big players than ever across video and social.
That’s exactly when you need independent architecture to keep things operationally viable, transparent and privacy-first. We’ve long spoken to the wrongs of marking your own homework, that a tech vendor can’t be the prosecution, the defence and the jury. For just as long we’ve had nods of understanding, but also of acceptance that it’s all too hard to create your own course beyond these walls.
Agentic AI has changed everything. When AI is making more of the everyday decisions, you want the platform it runs on to be independent of whoever’s selling the media. Independence also means your data stays yours: we don’t expose a client’s campaign data to outside AI models to train on, and that matters a lot more here as first-party data rules tighten.
It’s like a credit card network: you wouldn’t let the buyer’s bank or the store’s bank run the whole transaction and settle their own fees. You need a neutral party in the middle that routes it, formats the data so it works at any checkout, and gives everyone one trusted receipt. That’s the role we play in advertising, and it matters more as the buy and sell sides blur.
B&T: There’s understandable nervousness about AI making decisions with real media budgets. With NIVO, how much control does the advertiser actually keep?
GB: All of it, by design. NIVO AI agents don’t act on their own. At each step the marketer reviews what the agent has proposed and approves it before anything goes live, whether that’s a campaign build, a QA fix, or a decision tree. We’ve deliberately built narrow, task-specific agents rather than one all-knowing system, partly so they stay focused and don’t get presumptuous.
The bit people forget is that AI has no business context. Something off kilter might be intentional. NIVO won’t be across a business decision like a rebrand, and there are real-world events that change what you want from delivery. For example, during Covid, no model had the context for why toilet paper was flying off the shelves while plenty of brands wanted to pull their spend on it. That’s a call that stays with a person. It’s what human-in-the-loop means in practice: the agent proposes, a person signs off, and governance and brand safety stay with the team.
B&T: What are Australian marketers actually asking you about most right now?
GB: Two things, mostly. The first is “show me it’s real”. People have sat through a lot of AI demos built on mocked-up screens, so they want to see live agents working on the platform. Fair enough. And we can: these are active agents our own teams, brands and holdcos are already building with, not slideware.
The second is interoperability. Brands and agencies have already invested in their own AI and tools, and they don’t want to tear that up to work with us. They want partners who’ll connect with what they’ve built rather than lock them into one stack, and that’s what the Orchestrator, the connector, is for. People are past experimenting. They want to know what actually changes how their team works, and whether it’ll connect to the tools they already run on, Monday or Airtable, whatever’s in their stack, rather than become one more thing to log into.
B&T: We can’t let you go without asking: how does someone actually get their hands on NIVO AI? Give us the sixty-second version.
GB: Easy. If you’re running real complexity, high campaign volume, lots of variants, multiple markets, heavy QA or compliance, you’re exactly who we want to build with. We’re taking on design partners now, so it’s hands-on with our team rather than a sign-up page. The fastest way is to talk to your Innovid rep, or just message me. We love to start with the messy, high-stakes stuff, because that’s where you’ll see the difference fastest.
