Michael Levitz
August 14, 2025
Image credit: Sam Gallagher, Medium
ChatGPT, Perplexity, Google AI Overview / AI Mode, and Grok are quickly taking over the role of research and discovery in the B2B buying journey.
Forester reported “89% of B2B buyers have adopted generative AI, naming it one of the top sources of self-guided information in every phase of their buying process.”
Search Engine Land just named this the post-SEO world, and Chad S. White is calling it the AI Takeover of Search.
For B2B marketing teams, the strategy of using broad-based content to fill the top of the funnel has lost its impact. HubSpot, master and teacher of the old content marketing playbook, experienced a 36% traffic drop at the end of last year.
Three changes came together quickly to transform the purchase funnel and rewrite the content marketing playbook:
Traffic dropped across the board. But something else happened: conversion rates spiked for brands that had in-depth content about their products and solutions tailored to specific audiences and needs.
Ahrefs reported the impact on their funnel: “At least for Ahrefs, AI Search has the highest Conversion / Visit ratio. AI search visitors convert at a 23x higher rate than traditional organic search visitors for Ahrefs.”
SEMRush also reported game-changing conversion rates: “We have seen that the average AI search visitor (tracked to a non-Google search source like ChatGPT) is 4.4 times as valuable as the average visit from traditional organic search, based on conversion rate.”
This is creating a new math for content: lower total total traffic but higher conversion rates.
It's also creating a new model for content marketing in the era of AI search: fresh content positions a brand as the definitive authority and solution to current needs.
It's no longer enough to create content that matches a user's search query. Brands need to provide an answer.
ChatGPT, Perplexity, and co. consider themselves "answer engines." Not search engines. They're not competing take search traffic away from Google. They're competing to eliminate the need to search. By giving users answers.
Getting the high conversion rates from AI search comes with a new challenge: creating the low-funnel, answer-focused content that answer engines and prospects are looking for.
This requires ongoing calibration of a brand’s content strategy, continuously dialing it to the center of industry trends, social conversations, and search behavior. So that your brand is the substance and solution to the questions that matter to your sales cycle.
The goal is no longer to be visible in search results or get the click. AI search has erased that. The goal is to show searchers how you will accomplish their goal.
Traditional search was a means to an end. AI search jumps straight to delivering the end. Moving a prospect out of research and into action.
When you search for competitive intelligence on Google, you might search for a downloadable template or a blog post about how to create a competitive intelligence report.
When you search for competitive intelligence on an answer engine like ChatGPT, you ask it to draft the report for you. You naturally jump straight to doing the work.
As a content marketer in this new world, your goal is to create content that directly helps your users do their work.
For the past year, we’ve been working on exactly this challenge at Forecast.ing. I’d love to say we saw this coming, but we didn’t.
We got obsessed with the power of vector databases and semantic search. We realized we could capture substance and meaning over keywords to give brand marketers a competitive edge.
Doing this at scale wasn’t possible without AI agents, LLMs, Retrieval Augmented Generation (RAG), and machine learning. So we created an automated pipeline that connected the dots.
Then we booked a bunch of sales calls to gush about the power of unlocking the semantic meaning trapped in their data.
I'd recommend not trying this at home. Or at all.
We forgot to drink our own KoolAid. We weren't offering an answer, or helping our audience do their work. As one content marketer told us, "I was hoping you were going to make my job easier. But it feels like now I have more work to do."
That feedback became our north star. We had to go back and focus on the real Job to Be Done.
Our goal became simple: give brands a consistent voice in content marketing and demand generation, letting them communicate frequently, with a perspective unique to their brand, and the rigor of a always-on business intelligence team feeding them a reliable stream of fresh insights.
And we wanted this to be available to everyone with no setup.
We identified six critical data sources that, when combined, create a foundational view of market intelligence:
But connecting the data sources, constantly piping in fresh signals, and making sense of the data had to be automated, easy, done for you. Made for teams that are already lean, busy, and starved for time. We had to reduce their workload and increase their output.
We couldn’t give them another tool to configure, learn, and monitor.
Weeks turned into months of work as we simplified and made it set-it-and-forget it. We kept that quote front and center: "I was hoping you were going to make my job easier."
We dove headfirst into the hardest problems with no plan, no schedule. Just determination to come out the other side when the insights and content briefs magically appeared, ready to use.
When we finally came up for air, we had created an automated seven step process that runs once a week for each brand:
1. Agentic Web Retrieval: Our system intelligently gathers approximately 100,000+ content pieces from all six signal areas: your brand content, competitor blogs and newsletters, industry news, social conversations, search data, and uploaded documents.
2. Semantic Chunking: Every document gets broken down into contextually coherent chunks. Not just paragraphs, but meaningful pieces of information that preserve context and intent.
3. Vectorization & Storage: Each chunk gets converted into numerical vectors that capture semantic meaning, not just keywords. These vectors live in our specialized database where every piece of data can be traced back to its original source. No hallucinations, no mystery citations.
4. Topic Clustering: Using machine learning, we group related chunks together based on semantic similarity (cosine similarity) to identify core themes and emerging narratives across your entire market landscape.
5. Topic Modeling & Ranking: We analyze topic clusters to understand what's gaining momentum, assign descriptive labels, and rank topics by frequency, relevance, and recent activity. The system separates signal from noise automatically.
6. Brief Generation: For high-priority topics, the system generates actionable briefs with key insights, competitive positioning, audience language, powerful statistics, direct quotes, and content recommendations — all with citations to original sources.
7. Automated Refresh: This entire pipeline runs continuously. Every week, new insights and updated topic briefs appear in your dashboard. You never have to hunt for what's important. It comes to you.
We call this tool Brand Vector.
It’s more than a dataset; it’s a decision engine built for the new reality of marketing. The old playbook of chasing clicks is over. The new challenge is making your brand the definitive answer, the substance that AI search engines rely on to move a buyer from research into action.
Your new Job To Be Done (JTBD) as a marketer is to articulate why your brand is the best answer to a specific challenge right now.
Brand Vector constantly ingests and analyzes the content your brand, competitors, industry analysts, and audience are creating. Giving you a constant stream of brand-specific insights and briefs to activate.
All you need to do is log in.
