61% of your buyers already have an AI grading you

Your buyer is running an AI agent against you before you even know the deal exists. Today's issue covers what that means for your pipeline, a wild new outbound channel and a framework for proving a deal actually moved instead of just feeling like it did.

G2's 2026 Buyer Behavior Report finds 61% of B2B buyers already use AI agents in their purchase process and another 19% would for select use cases. The top agent use cases sit right in your funnel: comparing total cost of ownership, building vendor shortlists and evaluating the finalists. G2 also found buyers who let a chatbot build their shortlist bought from it 80% of the time, against 65% for buyers who didn't.

Why it matters for you: your deck isn't the first thing a buyer sees anymore, an AI's summary of your G2 reviews and pricing page is. If that summary is wrong or thin, you lose the shortlist before a rep ever gets a call.

A $127M-Funded Data Company Let an AI Agent Run Its Own Lead Qualification

ColdIQ founder Michael Lieben published a full breakdown of how Explorium (total funding: $127M since a 2021 Series C) runs its own GTM stack. CEO Omer Har's version: intent signals run ahead of the contact list, an AI agent qualifies every visitor and no human touches a lead until it's already routed to the right rep. Lieben posted the two-layer pattern worth copying on X: wire your intent data straight into a single coding agent instead of stitching five point tools by hand.

Cold Email Got So Good It Stopped Working. Now People Are Sending Cake.

Daymaker turns cold outreach into a physical object: you point at one person who matters, a bakery in that person's city bakes a custom cake and an operative hand-delivers it to the office. Your logo, your deck or your cease and desist joke goes on top. Daymaker reports roughly 30% same-day replies against the single digit reply rates a good email sequence earns. Its own results page cites a Sifted-covered VC campaign that landed 5 partner meetings from 7 delivered cakes. Sheel Mohnot's post about one delivery pulled over 120,000 views on X.

Why it matters: when every rep has the same AI writing the same personalized email, the scarce thing isn't relevance, it's being physically impossible to ignore. Pitch cakes run $99 flat and campaign cakes start near $100 plus delivery in SF, NY, LA and Berlin, so this is a play for the 5 accounts your sequence already burned, not a full list. The 30% and 100% open figures are Daymaker's own numbers, not third party audited.

Every sales leader knows the feeling. You walk into a pipeline review with a number you believe in, and twenty minutes later, you're defending every line item to a CEO who just wants to know what's actually going to close.

HubSpot Sales Hub ends that conversation. Every deal, every rep's activity, and every buyer signal are all in one place and updated automatically. So your forecast is built on what's actually happening. And when you present that number, you can stand behind it.

Get a demo

The Buyer-Evidence Ladder

If buyers are grading you with AI now (see today's Signal), your own pipeline needs the same rigor. Most CRMs still grade deals on what the rep did: demo sent, proposal sent, follow-up sent. None of that proves the buyer moved. The fix is a stage ladder where every rung requires a piece of buyer evidence, not a seller activity.

1. Discovery → Solution Fit: the buyer names the specific problem and its cost in their own words, in writing or on a recorded call. "They liked the demo" doesn't count.

2. Solution Fit → Business Case: a second stakeholder joins a call or thread unprompted by you. Single-threaded deals don't get to skip this rung no matter how warm the champion sounds.

3. Business Case → Proposal: the buyer shares a budget range, a procurement step or a target close date they didn't have to share. A rep-built ROI deck the buyer never reacted to doesn't count.

4. Proposal → Closed Won: legal or security review is scheduled with a named date. "They said next week" isn't a review, it's a hope.

The prompt: paste your open pipeline into Claude or ChatGPT and ask it to flag every deal marked past Discovery that's missing the buyer-evidence item for its stage. Bring that list, not a status update, to your next forecast call.

This isn't a new problem. RevOps teams have been rebuilding pipelines around it all year. Here's a clear writeup of why activity-based stages corrupt forecast data.

BEFORE: You define an ICP, hand it to a data vendor or three, then spend a week stitching Apollo exports with Clay enrichment and a scrubbed CSV before a single email goes out. By the time the list is clean the buying window has moved and half the contacts are stale.

AFTER: Landbase is an AI GTM platform: describe the audience you want in plain language and it returns a structured, verified, contact-level list built from its own dataset instead of a scraped one, then hands you messaging ready to launch. The company says campaigns that used to take months now launch in minutes, at roughly 80% lower cost.

Why now: Landbase closed a $30M Series A led by Ashton Kutcher's Sound Ventures and Picus Capital on top of a $12.5M seed, built by AppDirect co-founder Daniel Saks. If you're still paying for three separate data and enrichment tools to do what one prompt now does, this is worth a look. The 80% cost figure is Landbase's own claim from its site, not third party audited.

You’re in back-to-back demos. Granola is the AI notepad built for exactly that — jot what matters while it transcribes in the background and turns it into a clean summary with action items before your next call starts.

No meeting bots. Ask “What did I miss?” mid-call to get caught up instantly.

Try Granola free — code SALESBYTES for 1 month off any paid plan

To start getting immediate value from AI I've built the library (and will continue to build more) for top AI sales tools, prompts, tricks, and tips.