Latest / The Marketing Funnel Podcast with Fexingo: Awareness, Interest, Decision, and Action Stages / How HubSpot Rebuilt Its Marketing Funnel for AI
Transcript
- Lucas: Alright, so HubSpot. If you've been in marketing for more than five minutes, you know the name — they basically wrote the playbook on inbound marketing, the flywheel, all of that. But in the last eighteen months, they've done something pretty radical: they rebuilt their entire marketing funnel around artificial intelligence. Luna: Radical how? I mean, every software company is bolting on some AI feature. What makes HubSpot different? Lucas: Fair question. The difference is that HubSpot didn't just add a chatbot or a content generator. They fundamentally rethought how leads move through the funnel — awareness, interest, decision, action — and baked AI into each stage. They launched something called Breeze AI in mid-2025, and it's not one tool; it's a suite of AI agents that sit on top of their existing CRM. Luna: Breeze AI — I've seen that name. Isn't that their answer to Salesforce's Einstein? Lucas: Sort of, but the architecture is different. Einstein is mostly predictive scoring and recommendations. Breeze is more about conversational, real-time interaction. At the awareness stage, for example, HubSpot's AI can generate blog posts, social copy, and video scripts tailored to a specific buyer persona — and it does it in the voice of the brand, not generic GPT output. Luna: So the top of the funnel gets automated content creation. But doesn't that risk flooding the market with mediocre AI slop? Lucas: That's the concern, and HubSpot knows it. They've built in a human review step — the AI drafts, but a marketer approves. Still, the speed gain is massive. One HubSpot customer, a B2B SaaS company called Drift — no, wait, they were acquired by Salesloft — anyway, they reported that their content output tripled in the first quarter after implementing Breeze's Content Hub, with the same team size. Luna: Triple the output with no extra headcount. That changes the cost structure of the awareness stage. Lucas: Exactly. But the really interesting shift happens at the interest and decision stages. HubSpot introduced what they call 'conversational landing pages.' Instead of a static form asking for your email, you get an AI agent that asks qualifying questions in natural language — like a sales development rep, but automated. Luna: Oh, I've encountered those. I was researching marketing automation tools last month, landed on a HubSpot page, and this chat window popped up asking what my biggest challenge was. I typed 'lead scoring,' and it immediately sent me a case study about predictive lead scoring. It felt almost too seamless. Lucas: Right, and that's the key. The traditional marketing funnel assumes a linear path: you read a blog, you download an ebook, you get a call. HubSpot's new model compresses that. The AI can answer questions, send relevant content, and even schedule a demo — all in that first conversation. HubSpot's own data shows that conversational landing pages convert at two to three times the rate of static forms. Luna: Two to three times — that's not incremental. That's a step change. But does it actually lead to better qualified leads, or just more of them? Lucas: HubSpot claims both. They say the AI's qualification questions filter out tire-kickers early. So the leads that reach sales are more likely to convert. In their 2025 customer conference, they shared that one client — a mid-market cybersecurity firm — saw a 40 percent increase in demo to close rate after switching to conversational landing pages. Luna: Forty percent. That's the kind of number that makes a CMO sit up. But I wonder — does the AI handle objections well? I've had chatbots give me irrelevant answers when I push back on pricing. Lucas: That's the weak spot. The Breeze AI is trained on HubSpot's own CRM data, so it's better than a generic bot, but it's not perfect. HubSpot is transparent that the AI escalates to a human when it can't answer. The key is that the human gets the full transcript and context, so they don't have to start from scratch. Luna: So the handoff is smooth. That addresses one of the biggest friction points in the decision stage. Lucas: Exactly. Now, the action stage — the bottom of the funnel — is where HubSpot's AI really differentiates. They've introduced something called 'Smart Deals.' Instead of the sales rep manually updating the CRM, the AI listens to call recordings, reads email threads, and automatically populates next steps, risk flags, and even suggests the best discount to close. Luna: That's almost like having a virtual sales operations analyst on every deal. Lucas: Pretty much. HubSpot says that companies using Smart Deals see a 25 percent reduction in time spent on data entry, and a 15 percent increase in win rates because the AI surfaces insights the rep might have missed — like a competitor mention that requires a counter-strategy. Luna: Fifteen percent win rate improvement — that's massive for a B2B sales team. But here's my worry: if every vendor uses AI, doesn't the advantage disappear? It becomes table stakes. Lucas: That's the million-dollar question. HubSpot's bet is that their advantage lies in the data — they have over two hundred thousand customers' CRM data training their models. A smaller competitor can't replicate that. Plus, HubSpot is integrating AI across the entire funnel, not just one touchpoint. That holistic approach is harder to copy. Luna: So the network effect applies to AI too. The more customers use it, the smarter it gets. Lucas: Right. And HubSpot is also opening up Breeze AI to third-party developers. They announced a marketplace where you can build custom AI agents — for example, a real estate agent that automatically generates listing descriptions and qualifies buyers. That extends the funnel beyond traditional marketing use cases. Luna: That's smart. It turns the product into a platform. But let's go back to the funnel model itself. Does HubSpot still use the AIDA framework internally? Or have they abandoned it? Lucas: They've adapted it. In their internal documentation, they now talk about 'ai native funnels' where the stages are less sequential and more simultaneous. The idea is that the AI can engage a prospect at the awareness stage and, within the same interaction, move them to decision. The stages still exist, but the boundaries blur. Luna: That's a fundamental shift. The linear funnel has been the backbone of marketing for over a century. Are we saying it's obsolete? Lucas: Not obsolete, but the model is evolving. The classic AIDA assumed a one-way broadcast: you make noise, you generate interest, you persuade, you close. AI enables a two-way conversation at every stage. HubSpot's CMO, Kipp Bodnar, said in a recent interview that 'the future of marketing is conversational, not campaign-based.' That's a big departure. Luna: Campaign-based marketing — you launch a campaign, you measure results, you move on. Conversational marketing is always on. That changes how you budget, how you staff, how you measure ROI. Lucas: Exactly. And it changes the role of the marketer. Instead of writing copy and managing campaigns, the marketer becomes a curator and trainer — training the AI on brand voice, approving outputs, and analyzing the conversations the AI is having. HubSpot has already started retraining their own marketing team for this new model. Luna: If today's episode sparked an idea you've already started applying — or if you're thinking about how to adapt your own funnel — the reason we can keep having these conversations ad-free is listener support. It's at buy me a coffee dot com slash fexingo. Just a low-key way to keep this going. Lucas: Yeah, we love putting this together, and knowing it's useful is what drives us. So if you've gotten value, that's the place. Luna: Back to HubSpot — one thing I want to dig into is the pricing. How does HubSpot charge for these AI features? Is it a premium add-on, or is it included in the core subscription? Lucas: It's a mix. The basic Breeze AI features — like the content assistant and the conversational landing pages — are included in their Marketing Hub Pro and Enterprise tiers. But the advanced stuff, like Smart Deals and custom AI agents, require an additional per-seat fee. HubSpot calls it 'AI Credits' — you buy blocks of credits each month, similar to how you buy compute on AWS. Luna: So the more you use, the more you pay. That aligns their incentives with yours — they want you to get value, but they also benefit when you scale usage. Lucas: Right. And early adopters are seeing strong ROI. One HubSpot partner, a marketing agency called IMPACT, reported that their clients using Breeze AI saw an average 30 percent reduction in cost per lead within six months. That's a direct hit to the bottom line. Luna: Cost per lead down 30 percent. That's the kind of metric that gets a CFO to approve the upgrade. But I wonder — is there a risk of over-reliance? If the AI breaks or hallucinates, the whole funnel could stall. Lucas: That's a real concern. HubSpot has built in guardrails — the AI is restricted to certain actions, and all outbound communications require human approval. But if the AI misqualifies a lead or sends the wrong content, it could damage brand trust. HubSpot's CTO, Dharmesh Shah, has said they treat AI as an 'augmentation, not a replacement.' Still, the temptation to automate fully is strong. Luna: So the human in the loop remains critical. That's reassuring, but it also means marketers need new skills — prompt engineering, AI oversight, conversation design. Lucas: Exactly. HubSpot is investing heavily in training. They launched a free certification called 'AI for Marketers' that covers how to use Breeze effectively. Over 50,000 people have completed it already. That's a smart move — it builds loyalty and ensures customers get the most out of the product. Luna: So, bottom line: HubSpot is betting that AI will redefine the marketing funnel, and they're putting their own product at the center of that shift. For marketers, the takeaway is that the funnel isn't dead — but it's becoming fluid, conversational, and data-driven in real time. Lucas: Well put. And the question for every marketer listening is: are you ready to retrain your team and redesign your funnel for an ai first world? Because the competitors who do it first might not look back. Luna: And we'll be watching to see how it plays out. Thanks for joining us on The Marketing Funnel Podcast.