Episode Summary
In this episode of Content Logistics, host Baylee Gunnell sits down with Tom Rudnai, Founder & CEO at Demand Genius. They explore how B2B marketers should think about showing up in AI, and why citations alone miss most of the story.
Tom explains that prompts are not keywords, and that AI influence starts long before a buyer clicks a cited link. He breaks down the iceberg of AI visibility, warns against mass-producing content that creates content debt, and argues that positioning, content, and reputation must stay aligned.
He lays out a way to measure AI performance: track prompts across the journey, study sentiment and alignment, and focus on content with information gain. Tom closes by showing how AI can maintain content libraries while marketers spend more time on original research.
Guest Profile
Tom Rudnai
Key Insights
Citations only show the tip of the iceberg
Many teams treat AI like search and use citations as the main scorecard. That creates a narrow view of what buyers actually experience. In B2B journeys, citations often appear late, when a buyer already wants help making a decision. By that point, the model has already shaped how the buyer frames the problem, compares options, and builds requirements. That earlier influence is harder to see, but it matters more. A cited mention at the bottom of the funnel may reflect user intent, not true brand preference. That makes citations a weak north star on their own. A better approach tracks prompts across awareness, consideration, and conversion. It also looks at how the model describes the category, which brands it includes, and how often your strengths appear in the answer. Visibility matters. But deeper influence starts before a link ever shows up.
More content can create more confusion
Publishing more AI-written content can raise short-term visibility, but it also creates content debt. Every new page adds something the model can use to form a picture of your company. If that library includes old positioning, mixed messages, or thin pages built for prompt coverage, the model gets a messy view of who you are and who you serve. That hurts clarity. It also weakens your ability to show up for the right use cases. A useful way to think about this is through three connected levers: positioning, content, and reputation. Positioning defines who you are. Content communicates it. Reputation supports it through outside proof. When those three drift apart, AI picks up the gaps. Smaller teams should resist the urge to flood the site. Fewer, stronger pages that stay aligned over time will usually do more work than a large pile of disconnected content.
Original insight beats recycled summaries
AI is good at summarizing common knowledge. That changes the job of content marketing. If your content only repeats ideas that already exist, a model can produce the same answer in seconds. That means summary content has less value than it used to. The stronger play is information gain. Give the market something new to work with. That could be a fresh interpretation, new data, or a clear concept that helps people understand a problem in a better way. This kind of content does more than rank. It shapes the option pool that AI draws from when buyers narrow choices. Teams should also build prompt lists from real buyer language, not from their own website. Sales call transcripts, CRM data, and journey insights reveal the questions buyers actually ask. That gives you better prompts to track and better topics to create around.
Episode Highlights
Why AEO needs a different playbook
Timestamp: ~ 00:04:20
A common mistake is treating answer engine optimization like a simple extension of SEO. That framing keeps teams stuck in old habits. Search keywords are short and constrained. Prompts are longer, more personal, and shaped by role, context, and criteria. That changes how content needs to work. It also changes how marketers should think about optimization. Instead of matching one page to one phrase, teams need content that helps AI understand who they serve and what they solve. This section sets the tone for the whole episode. It shows why old search tactics still help, but no longer explain the full job.
“One is they equate a prompt and a keyword as if they’re the same thing. They’re really, really different. A keyword forces a user to consolidate their intent or question into a pretty short string of words. A prompt is totally different. It’s long, personal, it adds criteria, it assigns personas, it’s verbose.”
How AI narrows the field during a buying journey
Timestamp: ~ 00:19:50
Summary
AI does not work like a search page that fetches one answer and stops. It works through a conversation. Early in that conversation, the range of possible brands stays wide. As more context enters the exchange, the model applies criteria and narrows the set. That means strong visibility at the end of the process may simply mean a brand made it into the final pool. It does not always mean the brand out-optimized everyone else. This part of the discussion gives marketers a better model for how recommendation behavior forms. It also helps explain why upstream influence matters before direct recommendations appear.
“We take it as a problem, it talks to us, it applies criteria, understands who we are, may already have some of that knowledge, and then it gives us the answer. Over the course of that process, what it’s doing is applying criteria and narrowing down its option pool. Awareness stage prompts are super, super broad. By the end, it gets much more conservative.”
Use AI for maintenance, not just production
Timestamp: ~ 00:38:40
One practical use of AI sits far from content generation. It sits in content operations. Large content libraries drift fast as products, positioning, and claims change. Reviewing every page by hand takes too much time for a small team. This section focuses on using AI to monitor quality, flag outdated claims, and keep content aligned with current messaging. That kind of support frees marketers to spend more time on strategic work that still needs human judgment. It is a useful reminder that AI’s best role may be in the background, handling repetitive review work that most teams never have time to finish.
“AI does really boring, repetitive tasks really, really well, and it can use judgment. That job four years ago would’ve involved a poor human sitting down and literally every month reading every piece of content on your website and flagging, okay, this product description doesn’t match, this claim is out of date, this is no longer quite as relevant.”
The fastest way to spot lazy AI copy
Timestamp: ~ 00:41:10
The episode closes with a simple editorial test: can readers tell the copy came from AI at a glance? Repetitive comparison phrases, stiff structure, and overused punctuation make that easy. The point is not that AI should never help with drafting. The point is that weak review leaves obvious fingerprints. For teams under pressure to move fast, this is a useful reminder that clean editing still matters. Strong content should sound like a person who knows the subject, not a pattern machine repeating stock phrasing. It is a light moment near the end, but it lands on a real content standard.
“The one that has really started to wind me up is the ‘it’s not this, it’s that.’ The other ick is when you go onto a big business homepage and it’s just so obviously AI. The M dash is the other obvious one. Just go through it, turn it into a regular dash. That’s what ordinary people use.”
