Catch the Full Episode:
Overview
HubSpot spent over a decade teaching small businesses how to build inbound marketing funnels. Then AI-driven search cost the company 140 million visits in under a year, an 80% drop in traffic. Revenue kept climbing anyway.
John Jantsch talks with Kipp Bodnar about what replaced that lost traffic. They cover building a “taste profile” so AI tools produce work that sounds like you, the upside and risk of simulating customer reactions before spending a dollar, and why answer engine optimization means writing for machines that now read like humans.
This episode is for marketers, agency owners, and solopreneurs who’ve watched their own traffic slide and want a concrete next step versus another AI framework to file away.
Guest Bio
Kipp Bodnar is CMO at HubSpot, where he’s worked for over 15 years, joining when the company had under $10 million in revenue. He co-hosts the podcast Marketing Against the Grain and sits on the boards of Gusto and Similarweb. His new book, co-written with Kieran Flanagan, is Loop: Outlearn. Outmarket. Outgrow., releasing September 22, 2026.
Key Takeaways
- Build a taste profile (a few hours clarifying what you believe about your customers and your brand) before using AI for content. If writing it out is hard, have someone interview you and feed the recording to AI instead.
- AI buyer simulations get more predictable, but the trade-off is average results. Save your riskiest ideas for real-world tests instead of running everything through a simulation first.
- The 3 most-cited sources on tools like ChatGPT and Claude are currently YouTube, LinkedIn, and Reddit. If you’re not active there, you’re less likely to get mentioned.
- If you don’t publish your pricing, AI answer engines will guess it, and often get it wrong. Being transparent earlier in the buyer journey works in your favor.
- What separates strong content from AI sameness is what’s unique to you: your own data, your customers’ stories, your specific history.
Great Moments
- [01:36] – Bodnar explains HubSpot’s 140-million-visit traffic drop and the decision to diversify beyond the blog.
- [03:45] – The 4 steps of the Loop framework: express, personalize, amplify, evolve.
- [08:56] – The risk of over-relying on AI buyer simulations for predictable results.
- [12:37] – Answer engine optimization and the shift to writing for machines that now read like humans.
- [18:05] – The big question: is inbound marketing dead.
Memorable Quotes
- “My taste profile version is always way better. I’m normally a C with generic AI, and an A with a taste profile.” — Kipp Bodnar
- “When you’re trying to simulate customer reactions so your results are really predictable, the risk is you end up doing what everybody else is doing. You get a highly predictable result, but the magnitude is low or average.” — Kipp Bodnar
- “We spent the last 20 years writing for machines that were trying to parse information like humans. Now we have machines that act exactly like humans, so you need to write exactly like you’re writing to a human.” — Kipp Bodnar
- “Inbound marketing as we know it is dead because it was predicated on a world where information was scarce. Now knowledge is abundant, and actions and attention are the scarce things.” — Kipp Bodnar
- “When we saw our traffic drop that fast, it was like, there has to be a better way. If this is happening to us, I’m sure it’s happening to others, and we need to get ahead of it.” — Kipp Bodnar
Resources
AEO, AI content strategy, AI marketing, answer engine optimization, hubspot, inbound marketing, John Jantsch, Kipp Bodnar, Loop marketing book, marketing funnel, Small Business Marketing, taste profile
