Freemium isn't dead, AI search keeps getting weirder, and apparently CROs are building their own operating systems in Claude Code now.

This week's edition covers:

  • Kevin's Take: freemium isn't dead in the AI era, your finance team is just scared of the compute bill
  • The Signal: Signal: a live tracker of what frontier models actually cite, plus 24,000 prompts worth of SaaS visibility data
  • Tools & Tactics: five design decisions before you deploy a GTM agent, and what one CRO built solo in Claude Code
  • Quick Links: AI ad formats eyeing a quarter of US spend, ranking #1 on Google when the AI cites Reddit instead, tracking your brand across 10M prompts

Kevin's Take

Freemium isn't dead in the AI era, your finance team is just scared of the compute bill

There's a growing panic in B2B SaaS right now that freemium is dead in the AI era because every free user is burning real compute dollars. I think that panic is causing a lot of teams to make a really expensive mistake.

Elena Verna, who runs growth at Lovable, wrote a piece pushing back on this, and what got me is that Lovable is running a bottoms-up motion where most of their revenue comes through the free tier, and they're doing it profitably at scale, while every finance team in the valley is telling their CMO to gate the free product because the gross margins look scary on a spreadsheet.

I think what we're all missing is that gross margin at the user level is the wrong way to look at a freemium motion. It never was the right one honestly, but in the old SaaS world a free user cost you almost nothing to host, so nobody had to think about it. Now that each free session has a real Anthropic or OpenAI bill attached, CFOs are staring at a per-user cost and asking why we're giving this away.

The answer Elena makes really well is that you're not giving away product, you're buying distribution. Free is the acquisition channel. The math has to be done at the cohort and motion level, not at the individual user level, and if you kill the free tier to protect gross margin you're going to protect your way right out of your pipeline. The bottoms-up motion is what's actually driving the revenue, and you can't get the paid conversions without the free top of funnel feeding it.

One thing I've found useful when we get into conversations like this on our own work is to get in a room with the finance partner and reframe the whole thing before the budget cycle closes. Instead of defending the free tier as a cost center you're grudgingly keeping alive, we should show it as a paid acquisition channel with a CAC that includes compute, and compare that CAC to what we'd pay in ads or outbound to replace the pipeline it generates. My guess is the free tier wins that comparison almost every time, and it's a much harder number for finance to argue with than a vibes-based defense of freemium.

That's it for this week. Talk soon.

— Kevin Kerner, CEO, Mighty & True

The Signal

New tracker shows what frontier AI models actually cite (1 min read)

Latent Space built an original 'Frontier AEO Tracker' analyzing what sources frontier AI models (Astra and others) actually cite and choose, extending prior Claude Code research with new methodology. The tracker surfaces patterns across models and offers practical takeaways for improving AI answer-engine visibility.

Why it matters: If you're trying to get your brand cited by AI models, this tracker gives you a live, model-by-model view of what actually gets chosen so you can adjust your content strategy accordingly.

What 24,000 prompts reveal about AI search visibility for SaaS (1 min read)

An original study analyzing 24,067 AI search prompts reveals patterns in how SaaS companies get cited and surfaced in AI search results. The research offers a data-backed look at what drives AI search visibility for B2B software brands.

Why it matters: If you're guessing at your AI search strategy, this gives you real prompt-level data to benchmark your own SaaS visibility against.

Tools & Tactics

Five design decisions before you deploy a GTM agent (1 min read)

A GTM practitioner lays out five concrete design decisions marketing and sales teams must make when building AI agents into their go-to-market motion, including how to define what counts as an 'agent' versus automation. The piece frames agent design as a strategic choice set rather than a vendor feature checklist.

Why it matters: If your team is evaluating or building GTM agents, this gives you a decision framework before you get locked into a vendor's assumptions.

A CRO built his own revenue operating system in Claude Code (1 min read)

A CRO built an internal Claude Code-based operating system to run reporting, forecasting and coaching for a 100-person revenue org after just 8 months of learning the tool. The piece breaks down what he built and how other revenue leaders can replicate it.

Why it matters: If your GTM org still runs on spreadsheets and dashboards, this shows what a marketing-adjacent counterpart can build with Claude Code in under a year.

Resources & Tools

Strategy

How does your team compare?

The Mighty and True self-assessment compares your team's AI operating model, content velocity, and pipeline systems against mid-market B2B tech peers. Takes 10 minutes. You get a scorecard with concrete weak points you can attack this quarter.

▶ Take the Self-Assessment

Pipeline feel stuck?

Every week I talk to CMOs who are spending six figures on channels that stopped working two quarters ago. If your pipeline feels stuck and you are not sure which lever to pull next, let us talk. 30 minutes, no strings. Reply to this email.

▶ Book a strategy call →