Your buyers are sending AI agents to vet your stack, your attribution is drifting by device, and your creative tools just grew a brain.

This week's edition covers:

  • Kevin's Take: if your buyer's agent can't read your site, you're already out of the deal
  • The Signal: Signal: fresh benchmarks on AI budget gaps, plus the device and browser blind spots quietly breaking your attribution
  • Tools & Tactics: Canva's new native AI image model, and one marketer's actual agent org chart for his own job
  • Quick Links: a $3 ad testing workflow in Claude, three studies questioning the GEO hype, Salesforce buying Braze for $5.7B

Kevin's Take

If your buyer's agent can't read your site, you're not in the deal

The thing from Jason Lemkin that stopped me in his SaaStr piece on "ask your agent" is that his own AI VP of Revenue, which he calls 10K, is now the thing deciding which vendors get a real evaluation and which ones never make it to a human conversation at all. That is a very different world than the one most B2B marketing teams are built for, and I don't think we've really sat with what it means yet. We spent the last decade improving landing pages, nurture tracks, and sales decks for humans making a shortlist. If Lemkin is right, and I think he is directionally, the shortlist is now being put together by an agent that is reading your site, your docs, your pricing, your changelog, and your G2 reviews on behalf of a buyer who may never click a single link themselves.

The part I'd push a little further than the piece does is this. Lemkin frames it as a tracking problem, as in watch which competitors' agents are actually in production so you know who's winning. That's useful, but I think the more urgent question for us as marketers is whether our own surfaces are even readable to the agent doing the evaluation on the other side.

If an agent lands on your product page and has to figure out what you do from three rotating hero statements and a video, it is going to score you lower than the competitor whose docs clearly state the use case, the integration surface, the pricing model, and the API shape in plain structured text. That is not an SEO problem in new clothes. It is a different muscle, which is writing and structuring content for a reader that doesn't skim, doesn't get charmed by brand voice, and doesn't give you credit for the stuff you only said on a webinar.

What we're doing about this on our own site is going through every top-level page and asking a simple question, which is if an agent read only this page, would it be able to accurately describe who we help, how, and what it costs. In most cases the honest answer right now is no, and that is the work for the next quarter.

That's it for this week. Talk soon.

— Kevin Kerner, CEO, Mighty & True

The Signal

New data shows most CMOs still underfunding AI in 2026 budgets (1 min read)

LoudFace's 2026 report compiles benchmark data on B2B SaaS marketing and AI search, including that only 15.3% of marketing budgets go to AI initiatives while 56% of CMOs say they lack budget for their 2026 strategy.

Why it matters: These benchmarks give you hard numbers to justify or challenge your own AI budget allocation in planning conversations this quarter.

The hidden device and browser gaps breaking your attribution (1 min read)

This 2026 benchmarks report compiles the technical constraints reshaping marketing measurement, including ad-blocking rates, Safari's 7-day cookie cap, and iOS/Android split by market. It frames measurement accuracy as increasingly dependent on platform and device fragmentation rather than tooling alone.

Why it matters: Your attribution models are built on assumptions that are quietly breaking by market and device, and this report gives you the numbers to recalibrate before Q1 planning.

Tools & Tactics

Canva ships its own AI image model for all users (1 min read)

Canva launched Canva Craft, its first proprietary image generation model, rolling it out to all users with claims of better quality, more customization, and lower cost than prior third-party integrations.

Why it matters: If your team leans on Canva for campaign visuals, this gives you a native generation option that could cut your reliance on separate AI image tools and licensing costs.

One marketer mapped his own job into an AI agent org chart (1 min read)

A GTM AI automation lead at ServiceTitan shares a first-person account of how he mapped his own job into an 'agent org chart,' assigning AI agents to specific sub-tasks of his marketing role. The piece details the practical framework he used to decide which parts of his job to delegate to agents versus keep himself.

Why it matters: If you're trying to figure out where AI agents actually fit in your team's workflow, this gives you a concrete, practitioner-tested framework instead of another abstract maturity model.

  • Salesforce's $5.7B Braze acquisition reshapes martech vendor lock-in questions — Salesforce is acquiring customer engagement platform Braze for roughly $5.7 billion, consolidating martech capabilities and raising questions about vendor lock-in and AI-driven cam (via Ronan Shields)
  • Meta overhauls ad ranking with LLM-powered models — Meta announced three ad ranking upgrades: an Andromeda update now scoring every eligible ad per impression at over a billion ads per second, an upgraded GEM recommendation model, a (via <img src="https://www.adexchanger.com/wp-content/uploads/2026/07/34-120x120.png" alt="Photo of Allison Schiff" width="60" height="60" class="ai-author-image ai-author-image--circle ai-author-image--inline rounded-circle shadow-offset" loading="lazy"> Allison Schiff)

Strategy

Worth Reading

Bring us the problem. Leave with the plan.

The Blueprint is a free, 45-minute session with an expert advisor. Bring the goal you're behind on or the gap your team can't cover, and we'll come back with the fix and a 90-day plan to deliver it. No pitch attached, just an honest read on what is working and what isn't.

▶ Book your Blueprint

Help us with a research report

We are putting together a research report on how mid-market tech companies are actually allocating budget in 2026. Not the aspirational version, the real one. If you are willing to share your numbers anonymously, reply and I will send the survey. You will get the full report before anyone else.

▶ Share your numbers →