As someone who has been immersed in search marketing since the very early days, I’ve seen a lot of “new” ideas come and go. But every so often, something comes along that changes not just tactics, but the way we think about the economics of search itself. That’s what struck me in my conversation with Matt LeBaron, founder and CEO of Revvim, about their Ad AI platform and the broader future of paid search and AI-driven marketing.
Matt’s been doing paid search since 2004, with a deep foundation in SEO. He even wrote the original Wikipedia article on Quality Score back when most marketers were still trying to figure out how Google ads really worked. That blend of technical, historical, and practical experience comes through in how he thinks about one of the most deceptively simple questions in search: what should we do with branded queries?
Rethinking Branded Search and Incrementality
Most teams are familiar with the recurring debate around branded search: are we paying for clicks we’d get anyway via organic, or are we truly generating incremental value? The traditional way to answer that has been through holdout or incrementality tests. Turn off brand ads, watch what happens, argue over attribution, repeat.
Matt and his team approached this from a different angle: what if we architected technology that treated uncontested brand auctions differently, in real time, instead of relying on blunt, periodic tests? That thinking led to Ad AI, launched in 2020, which focuses on uncontested branded search (instances where your brand is being searched, you have the top organic listing, and there’s no competitor ad in the auction). In that environment, most advertisers are dramatically overpaying.
How Ad AI Turns Brand Waste into Savings
Ad AI monitors these conditions and, when it detects no competitor presence, swaps in lower-cost ad variants in close to real time. The response time averages eight minutes and forty-five seconds to unpredicted competitive changes in the SERP. In practice, that has meant about a 30% reduction in branded search spend for their clients and roughly $100 million in reclaimed budget across 350 enterprise and agency customers last year.
Interestingly, the platform’s strategy evolved. In the earliest days, they simply removed cannibalistic brand ads under the theory that organic would slide up and capture the click for free. That turned out to be too simplistic. Removing ads also removes sitelinks, promotional messaging, tailored landing pages, and, importantly, a lot of valuable analytics that are harder to get from organic since “keyword not provided” became the norm. The data have since shown that, on average, it’s better to keep the ad live 100% of the time, but at a far lower cost when you’re uncontested.
Navigating PMax and Shopping Campaigns
Things got more complex when Google introduced Performance Max (PMax). Early on, when Revvim removed ads, PMax would sometimes rush in with its own ad, undermining the strategy. That forced the team to innovate so their technology could always keep a manually structured campaign present and prioritized over PMax, as long as an exact match keyword existed.
From there, Ad AI expanded into Shopping, acting on both standard Shopping and PMax Shopping campaigns. When no competitor appears in the shopping block, the system can swap to lower-cost product ads, or, in some cases, full-price listings at lower bids for brands that don’t want to discount when they don’t have to. In those moments, brands can claw back margin while still reducing media costs.
Low-Cost, On-Demand Competitive Conquesting
One of the most intriguing extensions is Ad AI for Conquest. Instead of only defending and optimizing your own brand terms, you can have the platform monitor competitors’ brands and wait for moments when they’re not defending themselves.
In those windows, Ad AI automatically launches low-cost conquest ads, taking advantage of higher click-through rates and better quality scores that occur when you’re the only ad on the page. The result is something this industry hasn’t really seen before: conquesting that is both lower cost and effectively on demand. It also makes traditional brand holdout tests riskier, because any pause on your side can quickly become an opening for a well-armed competitor.
GEO, AI Search, and Setting Expectations
Zooming out, Matt and I also discussed the hype around AI search and GEO (generative engine optimization). Right now, AI-driven answer engines are mainly reshaping research intent, not the high-value engagement intent queries like “Nike” that still predominantly flow through traditional search interfaces. It’s rational to invest in visibility in AI environments, but we both believe marketers need to set realistic expectations with executives.
Discussion Points Include:
- How Ad AI identifies uncontested branded queries and reduces brand spend by about 30%
- Why removing brand ads often destroys value via lost sitelinks, messaging, landing pages, and analytics
- How Ad AI evolved to coexist with PMax and act on Shopping and PMax Shopping campaigns
- The mechanics and impact of Ad AI for Conquest and “on-demand” low-cost competitive bidding
- Why brand holdout tests may become less viable in a world of agentic AI watching your brand terms
- The distinction between engagement intent and research intent in AI vs. traditional search
- How GEO, PR, community engagement, and content velocity tie into future AI and search visibility