The whole game is changing because of agentic AI, which is completely altering how people find and buy things online. This means our old SEO playbook, stuffed with keywords and technical audits, is quickly becoming obsolete. You have to start thinking about how to influence an AI’s decision path. Our recent “Cognitive Commerce Ascent” campaign was designed to do just that, grabbing early market share for a niche B2B SaaS product by getting ahead of these AI-driven buying behaviors. The real question we’re all facing is, how do you get your products recommended, not just listed, by these new autonomous agents?
Key Takeaways
- Our “Cognitive Commerce Ascent” campaign delivered a 22% higher ROAS over previous efforts because we stopped chasing keywords and started feeding agentic AI content built for semantic relevance.
- We had to update our dynamic content schemas every two weeks. This was the only way to stay visible as the AI models kept changing how they understood product features.
- Instead of just listing features, our creative focused on problem-solution stories, which earned a 3.5% higher CTR in AI-curated content feeds.
- We moved 30% of the budget over to AI-driven content syndication platforms, and it paid off with an $85 Cost Per Lead (CPL), a huge drop from the $130 CPL we were seeing from our old search ads.
- By using direct feedback from AI agent logs to tweak content in real time, we managed to lift conversion rates by 1.8% in the middle of the campaign.
Campaign Teardown: Cognitive Commerce Ascent
Strategy: Anticipating Agentic Discovery Paths
The strategy behind “Cognitive Commerce Ascent” was based on one big assumption: by 2026, agentic AI wouldn’t just be fetching search results but would actively interpret user problems and recommend full solutions. This forced us to look past simple query matching. We bet that AI agents would favor content that showed a real grasp of a problem, had a clear value prop, and demonstrated authority. Our audience was small to medium-sized business owners in manufacturing who were looking for ways to automate their supply chain management. We set a budget of $150,000 to run the campaign over three months, from February to April 2026.
We ignored most long-tail keywords and instead built content around “semantic clusters”, groups of related ideas like “inventory discrepancies reduction” or “supply chain bottleneck identification.” This meant we had to map out the conceptual connections between these topics, trying to guess how an AI might link different bits of information to build a coherent recommendation for a user. It wasn’t a guess in the dark. A late-2025 eMarketer report showed that 45% of B2B purchase decisions were already being influenced by AI-generated summaries, so we knew we were on the right track.
Creative Approach: Solutions, Not Features
Creatively, we told stories that positioned our SaaS product as the clear answer to specific industry problems. We threw out the bullet-point feature lists. Instead, we developed content that presented a realistic business scenario and showed exactly how our product fixed the problem. For instance, we stopped saying “Automated inventory tracking” and started publishing pieces titled “How Manufacturer X slashed stockout costs by 18% using AI-driven reorder points.”
Our content mix included interactive case studies, short video explainers, and in-depth whitepapers. The video content was especially critical for getting noticed in AI discovery feeds, as it relied on strong visual storytelling and very tight problem-solution arcs. This is exactly where a specialist agency like Moburst came in handy. Their Video Production service was instrumental in creating the high-quality, conversion-focused videos that performed so well. They didn’t just make pretty videos. Their team understood the technical details of optimizing assets for AI ingestion, which is something a lot of creative shops miss completely.
Our creative team worked hand-in-hand with data scientists to figure out which emotional triggers and logical points were resonating with AI models trained on B2B buyer data. This meant constantly adjusting video scripts and headlines based on performance feedback, a much faster and more data-driven cycle than we were used to.
Targeting: Beyond Demographics
Of course we started with basic demographic and firmographic targeting, but the real work was in layering on what we called “intent-graph” targeting. This meant we were looking for digital footprints that showed a company was wrestling with the problems our product solves, even if they weren’t searching for our keywords. We were scouring industry forums, B2B news aggregators, and even patent filings to spot companies on the verge of needing a supply chain automation solution. Our main targeting parameters were:
- Company Size: 50-500 employees
- Industry: Manufacturing (NAICS codes 31-33)
- Technographic Data: Existing use of ERP systems (e.g., SAP, Oracle Netsuite)
- Behavioral Intent: Engagement with content related to “reducing operational expenditure,” “supply chain resilience,” or “predictive analytics for production.”
We put 40% of our ad spend on programmatic platforms that could handle this kind of complex intent targeting. The other 60% was divided between LinkedIn Ads for reaching specific job titles and Google Search Ads to catch the high-intent people who were already at the end of their buying journey.
What Worked: Semantic Depth and Visual Storytelling
The single most effective part of the “Cognitive Commerce Ascent” campaign was our focus on semantic depth. Any content that took a deep, exhaustive look at a single problem and provided a credible, detailed solution just crushed the more generic stuff. One interactive whitepaper, “The Hidden Costs of Manual Inventory: A Deep Dive into Manufacturing Profit Erosion,” pulled a 3.8% conversion rate from download to MQL. A standard “Benefits of Automated Inventory” blog post on the same topic only managed 1.5%. For these deep-dive pieces, our Cost Per Lead (CPL) averaged $85, which was well under our $100 target.
The video assets we developed with Moburst were also clear winners. We ran a series of 90-second animated explainers on B2B social platforms and various AI-powered content networks, and they hit an average Click-Through Rate (CTR) of 3.5%. By showing the problem visually before introducing our software as the fix, these videos did the heavy lifting for initial awareness. We ended up with 1.2 million impressions on those syndication platforms, beating our forecast by 15%.
Overall, the campaign’s Return on Ad Spend (ROAS) hit 3.2:1, a big jump from the 2.6:1 we’d seen on past campaigns that were more focused on keywords. The final cost per qualified demo request landed at $420, a 20% improvement from our baseline.
What Didn’t Work: Overly Technical Jargon
On the other hand, we found that content drowning in technical jargon with no business context fell completely flat. We saw that even sophisticated AI agents couldn’t bridge the gap between a term like “stochastic modeling for demand forecasting” and the actual business problem of “unpredictable sales cycles” unless we explicitly made that connection for them. Those articles had terrible engagement and high bounce rates. It was a good lesson: while an AI can parse complex data, its job in a buying context is often to simplify and find relevance. We had to go back and rewrite those pieces, always starting with the business pain point first.
Static infographics also flopped. They looked nice, but they just didn’t have the narrative flow to hold attention in an AI-curated feed. Their CTRs were stuck below 1%, and they produced almost no conversions. It was clear the AIs (and the humans watching) preferred content that told a story.
Optimization Steps Taken: Real-time Iteration
We made a few critical changes mid-campaign based on the data we were seeing. We pulled 15% of the budget from static display ads and pushed it into video syndication. We also built an “AI readability score” into our pre-publication checklist, using NLP tools to check our content for clarity and semantic density so it would be easy for both people and AI agents to digest. I’ve found that ignoring this step is a fundamental mistake. If an AI struggles to interpret your message, it simply won’t recommend it.
We also set up a tight feedback loop with the sales team. They would tell us what kinds of questions prospects were asking after seeing our AI-recommended content which let us spot content gaps instantly. We could then quickly produce articles or videos to answer those specific questions. For example, a pattern of questions about data security led us to publish a whitepaper on our platform’s ISO 27001 compliance which became one of our best-performing assets overnight.
Our content schemas, the structured data that explains our product to machines, were updated bi-weekly. This constant tuning made sure our product data was perfectly matched to how the AI models were evolving their own understanding of user intent. In my opinion, this kind of responsiveness is the single most critical factor for SEO success in the agentic AI era.
The “Cognitive Commerce Ascent” campaign proved that winning in the age of agentic AI means deeply understanding how these systems think and make recommendations. Marketers have to get beyond keywords and build rich, problem-solving content that anticipates what an AI is looking for. It’s a non-stop cycle of creating content, analyzing its performance, and iterating fast, all to make sure the autonomous agents guiding buyers choose your brand. For more on optimizing your content, check out these 5 shifts for 2026.
What is agentic AI in the context of consumer buying?
Agentic AI is basically an AI that acts on its own to help a user. It doesn’t just wait for a search query. It interprets a person’s complex needs, infers what they’re really trying to do, and then proactively suggests products or services. For buying, it means the AI is a personal shopper that understands context and makes independent recommendations, not just a search engine spitting out links.
How does agentic AI impact traditional SEO strategies?
It completely changes the game. Traditional SEO was about ranking for specific keywords. Now, you have to focus on semantic relevance and proving you can solve a user’s problem. AI-era SEO requires content that shows deep expertise and is structured so a machine can easily understand its value. You’re optimizing for concepts, not just words.
What is semantic depth and why is it important for agentic AI?
Semantic depth is how thoroughly your content covers a topic, including all the related ideas, nuances, and outcomes. It’s important because it gives an AI agent a complete picture. Content with good semantic depth is seen as more authoritative and accurate, making it far more likely that the AI will recommend it to a user with a complex problem.
Can video content be optimized for agentic AI?
Yes, absolutely. You optimize video for AI with a clear story, a quick problem-to-solution structure, and strong visuals. Behind the scenes, you also need to provide accurate transcripts, detailed descriptions, and structured metadata so the AI knows what the video is about. High production quality also helps because it drives engagement, which is a big signal of value for AI algorithms.
What are “intent-graph” targeting and dynamic content schemas?
“Intent-graph” targeting is about finding users based on complex patterns of behavior that signal their needs, rather than just what they type in a search box. A dynamic content schema is a machine-readable framework for your content’s data. It’s “dynamic” because you have to constantly update it to match new product features or shifts in how AI models interpret information, ensuring you stay visible.
