For two years, AI marketing tools have mostly been dashboards. Sharper data, richer charts, precise reports on exactly where you were losing… and no help closing the gap.
That era is ending. The agents have arrived, and they don’t watch the work. They do it.
Case in point: AirOps launched Quill, an AI agent that monitors your content, refreshes stale pages, drafts new ones, and even corrects outdated information about your brand on third-party sites. The early numbers they’re reporting:
- Asana saw a 93% increase in ChatGPT citations in two weeks
- Parallel saw 165% more citations
- Early customers report roughly 1.5x AI search citations and a near-50% lift in share of voice
Those are vendor numbers. You should be skeptical of vendor numbers. I was too.
So I went inside the tool myself and recorded exactly what I found… what it does well, where it falls short, and whether the hype matches the product.
Three things stood out while testing it, and they matter beyond this one tool:
- Content is becoming a system, not a project. Agents run continuously. A competitor makes a move, a citation slips, and the response starts before your Monday standup. Teams still operating on quarterly content calendars are bringing a schedule to a race.
- The third-party correction piece is the sleeper feature. Nearly 90% of AI citations for B2B SaaS come from off-site sources. An agent that fixes wrong information about you on other people’s websites is working on the surface that actually decides your AI visibility.
- The human review gate is where quality lives. Quill flags actions for approval before publishing. That gate is the difference between an agent that compounds your brand and one that floods the internet with slop wearing your logo.
Meanwhile… last week’s story gets its verdict tonight.
Reddit reports Q2 earnings after the close today, and options markets are pricing a swing of roughly 12% in either direction. The Google standoff we covered last week is the reason. Steve Huffman has publicly signaled a shift from flat-fee licensing (~$60M/year from Google, ~$70M/year from OpenAI) to usage-based pricing on 2027 renewals. That WSJ story about cutting Google off? That was the opening bid.
My read stands from last week: whatever tonight’s print says, human conversation data is being repriced in real time. The platforms holding authentic human content are done renting it out cheap… and the marketers who understand that shift are positioned for what comes next.
Create once. Distribute forever.
— Ross
📊 The Data Point
Wells Fargo estimates Reddit could be walking away from as much as $500M in AI licensing revenue if the Google standoff goes badly. That’s the size of the bet Reddit is making that its data is underpriced. When a platform bets half a billion dollars on the value of human conversation, believe them.
⚡ Quick Hits
- AI search is projected to overtake Google search by 2030, per researchers tracking the generative shift. The “10 blue links” era has an expiry date.
- Your brand is being described by AI right now… and some of it is wrong. On Thursday, August 6 at 12pm ET, I’m co-hosting a live session with Profound’s Nick Lafferty on finding those errors and correcting them at the source. Save your spot
- New on the Lab: most brands show up on Reddit sounding like a press release. Our new piece breaks down how a branded persona account with actual personality unlocks what corporate accounts can’t.
🔗 From the Lab
- Personality on Reddit Matters More Than You Think: How a Branded Persona Account Unlocks Real Brand Possibilities
- Nearly 90% of AI Citations for B2B SaaS Come from Off-Site Sources
- How Vanta’s $517K/Month SEO Moat Translates Into AI Visibility
Want to know how visible your brand is inside LLMs? Just reply.
