If AI agents are already shortlisting vendors and scoring signals keeps misfiring, your stack needs a rethink before the next pipeline review.

This week's edition covers:

  • Kevin's Take: your site still writes for humans, but the agents are already shopping it
  • The Signal: Signal: why the same intent signal predicts deals at one company and noise at another, plus a new benchmark for which B2B brands AI engines actually name
  • Tools & Tactics: a practitioner's playbook for sponsoring B2B creators, and how Resend got AI coding agents to recommend it over bigger rivals
  • Quick Links: YouTube out-citing your own site in AI answers, ChatGPT ad traffic converting at a third of Google's rate, visitor ID accuracy under the microscope

Kevin's Take

The agents are already shopping. Your site still thinks a human is reading it.

About two years ago I would have told you that the job of a B2B website was to help a human make a shortlist. We've been testing what happens when you let an agent do a vendor evaluation from start to finish, and what keeps jumping out is how badly most B2B sites perform when the reader isn't a person. The nav doesn't matter. The hero animation doesn't matter. The gated PDF is actively hostile. What the agent wants is structured answers to specific questions, and if it can't find them in about two passes it moves on and recommends somebody else. That's what Olivia Moore's piece on the B2A economy gets right, and I think a lot of CMOs are going to under-react to it, because the buyer in a lot of our funnels is no longer a human at all, it's an agent acting on behalf of one, and the whole stack we've built over the last ten years is tuned for the wrong reader.

Where I'd push the a16z framing a bit is that I don't think "competing for agents" is really a marketing problem the way the piece sets it up. It's a product and content-architecture problem that marketing is going to get handed because nobody else owns it. Agents don't care about your brand campaign. They care about whether your pricing is on the page in a form they can parse, whether your integrations are listed somewhere other than a logo wall, whether your docs answer the actual question a buyer is trying to answer, and whether your claims are specific enough to be verified against a competitor in the same query. That is a very different brief than the one most content teams are working from right now, and the teams we see figuring this out first are quietly getting included in shortlists they never knew they were being considered for.

One thing I've found useful is to open a fresh browser, give an agent a real buying prompt in your category, and watch what it does with your site compared to the three competitors it also pulls up. If you come out of that watching the agent skip past you on the second question because your answer wasn't on the page, you've found your Q1 roadmap, and it probably isn't the one in your current plan.

That's it for this week. Talk soon.

— Kevin Kerner, CEO, Mighty & True

The Signal

Why the same buying signal predicts deals at one company and nothing at another (1 min read)

Upside analyzed 102 buying signals across 19,459 deals and 205,036 contacts at ten B2B companies, finding the same signal can produce a 15x lift in conversion at one company and zero predictive value at another. The research argues that generic buying-signal scoring models don't transfer across companies without being recalibrated on each company's own deal data.

Why it matters: If your team is buying a generic intent or signal product off the shelf, this data suggests you're scoring leads on patterns that may not actually predict buying behavior inside your own pipeline.

New benchmark ranks which B2B SaaS brands AI engines actually recommend (1 min read)

A new benchmark dataset analyzes 94 B2B SaaS brand entries across five G2 categories to measure which companies AI answer engines cite when buyers ask for software recommendations.

Why it matters: If you're not tracking whether AI engines mention your brand by name, you're flying blind on a channel that's already shaping buyer shortlists before they ever hit your site.

Tools & Tactics

A practitioner's playbook for sponsoring B2B creators (1 min read)

Emily Kramer shares first-person advice on sponsoring B2B creators, drawn from working with 60+ brands as sponsors of her own newsletter. She covers why creators work, when to use them, and how to structure sponsorships for maximum effect.

Why it matters: If you're weighing creator sponsorships as a channel, this gives you a practitioner's playbook from someone who's sat on both sides of the deal.

How Resend got AI coding agents to recommend it over bigger rivals (1 min read)

Latent Space interviews Resend CEO Zeno Rocha on how his team engineered docs and tooling so AI coding agents like Claude Code recommend Resend over larger incumbents like Cloudflare Email. The conversation frames this as 'designing for agents' — a new layer beyond traditional AEO/SEO.

Why it matters: If AI agents are now the ones recommending vendors to your buyers, your docs and product surfaces need to be optimized for machine readers, not just humans and search crawlers.

Listen: Tech Marketing Rewired

Tech Marketing Rewired is the podcast for CMOs rebuilding their orgs for AI. Each episode is an unscripted conversation with a working operator on what they're actually shipping, what broke, and what they would do differently.

▶ Listen to Tech Marketing Rewired

Forward this to one person

If you find Drag and Drop useful, the best thing you can do is forward it to one marketing leader who needs it. Not your whole team. Just the one person who would actually read it. That is how we grow, and it is how they discover us.