AI chat assistants are primarily used for cognitive tasks
AI chat tools are not being used the way marketers hoped. Most executives still think of them as assistants geared toward closing sales or delivering ads more efficiently. The problem is that the data doesn’t back that up. Real-world use shows these systems are helping people think, write, analyze, and plan. They’re not selling, at least not yet.
A broad analysis led by Dan Petrovic, Director at AI SEO agency Dejan, looked at 3.9 million conversational turns, comprising over 613 million words. The result? Nearly 65% of these interactions had zero commercial intent. People are using AI to get work done, not to buy shoes or sign up for services. The most common uses include brainstorming (7.7%), planning (6.5%), and even having emotionally supportive conversations (6.2%).
For leadership teams, this changes how you should think about AI in your go-to-market stack. If your product or content strategy is built mainly around selling, meaning it’s only focused on driving immediate conversion, you’re going to miss where people are actually showing up. AI is being used to solve problems before customers even know what they need to buy, which changes everything about how and when you make your value visible.
AI chat sessions are short, fast, and task-oriented
The average AI chat isn’t long. That’s not how people are using these systems. Everyone wants to move fast. The analysis from Dejan found that most AI chats consist of just two conversational turns, a user’s request and the AI’s response. Median session length? 430 words. Over 80% of them stay below 1,000 words.
These aren’t drawn-out support conversations or in-depth product evaluations. They’re quick, focused inputs with an expectation of equally sharp output. The AI isn’t there to hold a discussion, it’s there to finish a task.
This creates an important insight for business leaders: brevity matters. Your AI optimization strategy can’t just dump long, keyword-heavy content into these systems and expect users to engage. The AI is doing most of the work, generating roughly 83–84% of the content in any given exchange. People are leaning on it as a tool to produce immediate results, whether that’s drafting a summary, outlining a proposal, or analyzing data on the fly.
So, if your business operates in a space where fast answers matter, or where efficiency is a differentiator, you need to structure your content for AI use, not just web search. Otherwise, it’s going to be ignored.
Commercial intent in AI use is early-stage and limited
AI chat interactions are not driving purchases. The assumption that users are plugging into chat assistants to buy things doesn’t hold up. Dan Petrovic’s analysis of more than 24,000 categorized sessions showed that only 35.4% had any commercial intent at all. And within that small piece of the pie, most were early in the funnel: 10% were awareness-driven, 8.5% were in consideration, and a combined 6.9% fell into discovery and decision support.
Actual transactional activity, like making a purchase or completing a form, was seen in only 4.8% of sessions. Post-purchase use, such as help with product setup or troubleshooting, was slightly higher at 5.1%. So, even in the commercial segment, it’s more about learning, not buying.
This matters if you’re making budget decisions around marketing, sales, or content alignment. Your AI-facing strategy should not be built around conversion. It needs to serve the informational and exploratory needs of users who are just figuring out what they care about. If those early conversations feel useful, if the AI points them to your insights, they’ll remember your brand when they’re ready to act. But the content has to show up early, and it has to be built for discovery, not push.
AI conversations function as cognitive workflows
AI assistants aren’t working the same way as search engines. Most SEOs still think in terms of keywords, hoping to match what someone types with the copy on their landing page. That approach doesn’t translate to AI usage. These systems don’t just find content, they help build responses in context. That shifts the value model completely.
What Dan Petrovic found is that AI is contributing the majority of the content in these conversations, about 83–84% on average. This shows users are sharing minimal input and relying on the assistant to expand, organize, and output the result. That’s not how traditional search engines operate. It’s a multi-step process where the assistant becomes part of the work.
From a strategic standpoint, this is where most teams need to shift. If your content isn’t structured for input reuse, if it doesn’t help the AI formulate better answers over multiple steps, it won’t surface in useful ways. You need to stop optimizing solely for individual keywords and start designing your thought leadership to function as modular, high-context input the assistant can reassemble.
This means building out content that helps users work through complexity, not just delivering surface-level answers. When your material can support tasks across stages, writing, summarizing, planning, it stays in circulation, carried forward by the AI through multiple user intents. That’s real value. And it’s where visibility happens at scale.
AI content strategy should prioritize cognitive support
Most current AI content strategies are misaligned. They focus heavily on transactional keywords, designed to close deals or funnel users toward purchases. That approach is disconnected from how people are actually using AI. The usage data is clear, people engage with AI assistants to think better, not to buy faster. The focus is on cognitive utility: learning, planning, analyzing, writing.
Dan Petrovic, Director at Dejan, analyzed millions of real-world AI interactions and confirmed what most strategies haven’t accounted for, AI doesn’t act as a sales agent, it acts as a co-pilot in thought. That means what shows up in discovery isn’t determined by what converts the fastest. It’s shaped by what helps the assistant perform cognitive tasks.
If your content can support a workflow, help someone write a summary, evaluate a framework, plan a timeline, it gets reused. And each reuse increases exposure. That’s not about branding. It’s about utility. Executives who want to compete in AI environments must audit whether their content improves decision-making and task execution, not just whether it pushes users toward checkout.
It’s time to build for relevance within workflows. Structured, modular information that fits cognitive usage patterns will outperform keyword-dense pages built for SEO ranking. If it helps the AI complete the user’s task, it stays visible. That’s the real metric that matters now.
Main highlights
- AI drives thinking: Executives should shift their AI content strategy to support cognitive tasks like planning, writing, and analysis, as 64.6% of sessions show no commercial intent.
- Short tasks dominate AI usage: Most AI interactions are brief and goal-driven, leaders should prioritize fast-access, high-utility content optimized for quick execution.
- Commercial interest is early-stage: With transactional intent present in less than 5% of interactions, teams should invest in content that supports customer discovery and pre-purchase education rather than direct sales.
- AI behaves more like a collaborator than a search engine: Instead of keyword matching, content needs to support multi-step workflows, organizations should focus on structured, reusable assets that enhance task completion.
- Content should enable outcomes: To stay visible in AI environments, teams must build context-rich content that helps users think and act, not just convert, this is where long-term brand value and engagement grow.
