Presenters

Source

Meet Your Newest, Most Powerful (and Uncontrollable) Team Member: The AI Product Manager! 🚀

Hey tech enthusiasts and business leaders! Pavel Fabrikantov, from Semrush, here. Today, we’re diving into a topic that’s fundamentally reshaping how businesses acquire and retain customers: the rise of AI search and its impact on your product’s journey. Forget everything you thought you knew about the traditional marketing and product funnels, because AI has rewritten the rules.

The Great Search Shift: From Weeks to Hours ⏳➡️⚡

Remember the old days? When a business decision meant sifting through ten Google links, cross-referencing with Reddit, and chatting with colleagues? This process could drag on for weeks, sometimes even months.

Welcome to the new world. A user asks an AI, the AI picks a brand, and the user confirms, perhaps with another AI interaction, and makes a purchase. This entire journey can now take mere hours. The middle of the product funnel? It’s practically vanished.

This isn’t a future prediction; it’s a present reality. AI search traffic has exploded, growing 5x year-over-year, and it’s just the beginning. A staggering 80% of B2B buyers in tech now use AI as much as traditional search when researching new vendors. The catch? This traffic doesn’t appear on your standard dashboards. It’s the mysterious “direct traffic” that shapes deals before you even know it’s happening.

The Invisible Buyer Journey 🕵️‍♀️

Imagine this: someone, late at night, in a bar, looking for a new data management vendor. They’re using AI to find solutions before they even consider engaging with your sales team. This interaction happens before your beautiful onboarding pages, your premium tools, or your landing pages. AI has become the first point of contact, and you might not even be there.

The Myth of “One AI” 🤯

Here’s where things get really interesting. When we say “AI,” you might picture a single, monolithic entity – your favorite chatbot, perhaps. But that’s a misconception. There are multiple AI systems out there, and they all behave differently. Crucially, they have different opinions about your value proposition and your product.

Output Variations: More Than Just Links 🔗

Consider the difference in outputs between ChatGPT (often with links) and Google’s AI mode (sometimes just a plain text summary). But the nuances go deeper. The choice of LLM, like the difference between a powerful model and a more budget-friendly one like GPT-4o mini (which is surprisingly prevalent due to its cost-effectiveness), can drastically alter how your product is presented.

Source Matters: Feeding the AI Beast 🍽️

The sources that feed these LLMs during their pre-training phase are critical. A comparison of Walmart.com’s visibility across Gemini and ChatGPT showed similar visibility scores. However, the number of cited pages varied dramatically: 200,000 versus 1.6 million. This means different primary sources lead to different answers about your brand and your product.

The Shifting LLM Landscape: Gemini on the Rise? 📈

And for those who thought ChatGPT was the undisputed king, think again. While it grew steadily for years, its growth has flattened and started falling. Meanwhile, Gemini is going up. The AI market is not monolithic; each LLM is carving its own path, and who wins is decided independently.

Auditing Per Model: Your New Marketing Mandate 🎯

So, what do you do? The answer is simple, yet profound: audit per model, not per AI as a single entity. Test with four to five key platforms and 50 queries. Track mentions, accuracy, and who emerges victorious in the AI search results.

This isn’t just a discovery problem; it’s a decision-making challenge. You now have a team member you know almost nothing about and have minimal control over. AI is your first product manager.

The “Shadow PM”: Uncontrollable but Undeniably Influential 👤

In the past, product managers conducted customer interviews, built roadmaps, and prioritized features. Their work directly impacted what users saw first. Today, there’s a second PM that your division doesn’t control. We call it the Shadow PM. You didn’t hire it, you can’t fire it, but it’s working millions of times a day.

The Benefits of Managing Your Shadow PM 🌟

When AI sets the right expectations, users arrive aligned. If a user asks about 7-day delivery and the AI accurately confirms “yes, plus free returns,” they’re more likely to click through, find what they need, convert, and have a higher Lifetime Value (LTV).

How to Manage Your Shadow PM: Three Pillars 🛠️

  1. Prompt Research: Understand What’s Being Asked 🗣️

    • This is your new, scaled customer development process. Instead of ten interviews a day, you can analyze thousands.
    • Identify which prompts mention your brand, what AI responses look like, and where your competitors stand.
    • Key Insight: AI can only recommend your product if it can read and explain it. If it can’t crawl your product, it can’t recommend it.
    • Website Challenge: Many AI crawlers don’t render JavaScript. If your landing pages are heavily JavaScript-dependent, your AI visibility can drop to zero, regardless of your content quality. This impacts not just new users but also retention and LTV.
    • Documentation is Key: When users have problems, they’re taking screenshots and asking AI for help. If your docs are clear and crawlable, the problem is solved. If not, users blame your product and churn. Your documentation, help centers, and FAQs are now training data directly affecting churn.
    • Actionable Tips: Make your content extractable, unblock AI crawlers, server-render everything, avoid JavaScript walls, and structure your docs in short, answer-first chunks. Sites with structured data see up to 30% higher citation rates in AI.

  2. Sources: Amplifying Your Reach 🌐

    • AI doesn’t just affect acquisition; it can amplify expansion through third-party sources that feed AI.
    • When a user asks, “Can my tool do X?”, AI can respond, “Yes, and you can also do Y and Z,” leading to feature discovery and expansion without a sales team.
    • The Challenge: Your positioning isn’t fixed. A study of 200,000 prompts revealed that sites gained share on one platform while losing on another with the same content.
    • Earned Media is King: 77% of brand citations in AI come from third parties like Reddit, G2, LinkedIn, and industry reviews. Don’t wait for these platforms to go behind paywalls or strike exclusive deals with LLMs. Let these sources champion your value proposition and showcase additional use cases.

  3. Processes: The Growth Engine Loop 🔄

    • Combine prompt research and source optimization to shape AI perception.
    • Fix documentation, structure content, and ensure accurate AI recommendations.
    • This leads to users arriving with the right expectations, better activation, and improved retention.
    • Happy users discover new features, expand their usage, and share their success stories on platforms like Reddit and LinkedIn.
    • This user signal feeds AI, creating an even stronger loop. This is no longer just a marketing campaign; it’s a growth engine.

Data-Backed Growth: The P&L Impact 💰

AI-referred visits have a 27% lower bounce rate, are 12% more engaged, and are 5% more likely to convert. This isn’t due to tricks; it’s because they come with the right expectations. Imagine the impact of these numbers on your P&L and unit economics!

The AI Revolution: Multiplying Perception ✨

To put it in historical context:

  • Mobile multiplied distribution.
  • Cloud multiplied scale.
  • Marketplaces multiplied exposure.
  • AI engines multiply brand and product perception.

They amplify how your product is described, who gets shortlisted, which features matter, and which don’t. When you design for this system, AI truly becomes a growth engine, and you can finally gain control over your “Shadow PM.”

Thank you!

Appendix