Here is another demo of Docket, this time, a quick peek into the latest release - Conversational AI Sales Engineer for B2B Sales! Docket started as an answering engine. Like how an AE can ask questions to a sales engineer, they could ask questions to Docket and get an accurate

You can find the podcast here:

The math is brutal: You're paying $45+ per lead, scoring them with outdated rules, then watching SDRs burn through garbage "opportunities" that convert at 2-3%. That's not marketing. That's expensive hope.

This is a snippet from the conversation I've had with Steve on my podcast. Steve spent 7 years at Google and now runs his own agency, which is why he understands this game better than anyone else.

Me: That makes so much sense... Steve: The goal should be building a database, not trying to fill a funnel with fake opportunities.

Take it through an account-based lens where you're looking at - the account, - the people within the account, - their behavior, - their demographics, - their firmographics. Get smarter about who you want to talk to, when, and what you say.

Me: So what's the alternative? Steve: Reframe it. If you're buying syndication leads, you're not buying opportunities. Instead, you're buying a database that you can actually market intelligently.

Steve: You're lucky if you can get 2-3% of content syndication leads to turn into opportunities. You're essentially flooding the system with inefficiency by investing in humans to call in and qualify a massive volume of low-quality leads that marketing is spending a lot of money

Steve: That's the problem. Marketing talks to a lead gen vendor who says, "I'll give you 10,000 leads at $45 a pop, put them in your nurture, and those will turn into opportunities." Marketers think this is the easiest ROI. Me: And it's not?

Steve: If you are buying leads, scoring them with a dated system, and then sending them to sales, I can almost guarantee right now 50% of that activity is inefficiently wasted budget. Me: But everyone's doing it 👀

Me: What's the biggest inefficiency you see in go-to-market today? Steve Armenti: The whole lead-to-MQL model is done. Me: Wait, what? That's like the foundation of most marketing teams. 🧵

@HubSpot @INBOUND If you're attending INBOUND tomorrow, please feel free to stop by booth 15 (located next to the Tech Stack Showcase stage).

@HubSpot @INBOUND 5/ We also realised that attendees are struggling to find info about INBOUND. So our engineering team pulled an all-nighter & cooked up the Docket AI Concierge. Think of it as a personal event assistant to find sessions, halls, and answers instantly -

@HubSpot @INBOUND 4/ Marketing leaders are facing significant board pressure to implement AI to improve their performance, which is leading to poor decisions as leaders are forced to check boxes without nuance.

@HubSpot @INBOUND 3/ Entry-level jobs, especially in content generation, will be more challenging. If you still want to enter the content paradigm, please learn to use AI as fast as you can, or you'll become unemployable.

@HubSpot @INBOUND 2/ AI is far from replacing humans, and most martech/salestech tools have also rolled back from replacing their positioning from "replacing humans" to "expanding human potential."

@HubSpot @INBOUND 1/ People are confused about AI, and many organisations are still figuring out what it means to have practical AI.

We spoke to ~280 prospects, printed hundreds of superhero cards, and wrapped up Day 2 at @HubSpot's @INBOUND 2025. Here are some key observations 🧵

That's pretty much it Don't complicate B2B sales & marketing

4/ Do you have 1 - 3 sustainable channels where you can reach prospects, or where prospects can find you?

3/ Is your messaging bridging the gap between what your product solves and the promised land for your ICP?