Josh Grant

Josh GrantSep 15, 2026

The first time I met James Cadwallader almost two years ago, we hit it off over a shared obsession: you have to keep reinventing your company and re-earning your place in the market. Every few months. Even when things are working. Today, Profound announced a $180M Series D co-led by Sequoia Capital and Kleiner Perkins. That belief is a big part of why. I've worked with dozens of people at Profound since that first conversation (shout out to Nick Lafferty, Eleanor Hooker, Trevor Pyle to name a few) and spent time with the team in their NYC office. Getting to know the people behind the product has been one of the best parts of the last two years. AEO gave marketing teams a new set of questions to answer. Where and how are we showing up? What sources are driving citations? Why is a competitor getting recommended? But knowing the answer still leaves someone on your team responsible for figuring out what to do, briefing the work, chasing it down, and getting it shipped. That gap is where I'd expect the next generation of marketing platforms to get built. Profound's AI Marketer, Aim, is a move in that direction. It uses your brand's context, answer-engine data, and connected tools to identify opportunities, recommend a plan, and execute work your team approves. That has implications well beyond AEO. Brand, PR, PMM, growth. Every function has work that stalls between "we should do this" and someone actually doing it. I want marketing teams spending more of their time deciding what's worth doing, developing ideas, and making the work better. Giving them the capacity to act on those decisions is the part of AI in marketing I'm most excited about. James and team: congratulations. You've kept earning it.

Josh GrantSep 11, 2026

Case study: leave your job. Four months full time on StackedGTM. Seven figures booked in the first quarter. I'm not chasing any of it. It's what happens when every hour goes to work that matters and nothing else. The revenue isn't what I'd tell you about over a beer. This is: I've always built. At Affirm it was campaigns and the systems behind them. At Webflow it was the growth machine itself. But the ratio has flipped. In-house, I fought for time to build. Now it's most of my day, and all of it is for clients. Lovable for the apps I build inside client portfolios. Claude for the thinking and the code behind them. Supabase holding the data. Profound for the AEO and marketing agent work, as that's where most of my clients are trying to win right now. Gamma when a client needs to see it. And a newsletter on beehiiv that 21,000+ people trust to teach them no frills AI GTM and AEO. The gap between having an idea and having it live for a client is now hours. And the loop that's formed around it has been the wildest part. I build for a client and the business grows. I write about what worked and what didn't. The following grows because the wins are real and so are the failures. And the writing brings the next client in. Build, write, grow, repeat. I didn't design it. It just started running, and every month it runs a little faster. If you've been in a leadership seat long enough that building became a luxury, take it back. What I've learned: - You learn as fast as the number of bets you get to watch. In-house, you run one plan at one company and find out later if it worked. Now I'm inside a dozen-plus companies at once, at different stages, making different calls, and I watch them play out every week. Dozens of bets running at the same time will always teach you more than one running after another. That's not a knock on in-house. Nobody told me that, and it took leaving to see it. - Building in the open is the whole pipeline. I braced for a fight against the noise of bad AEO vendors. It never came. Nearly every engagement has come through a founder, an investor placing me across their portfolio, or a newsletter someone forwarded. I still haven't made a deck. - Admin is a real job. Quotes, follow-ups, AP, renewals, contracts. I underestimated it badly and I've automated as much as I can (side note: I'm 48 hours into using Instinct as a virtual EA and my mind is blown) - The people. I spend my days with some of the sharpest people building in AI, at companies I would have paid to just sit near a few years ago. I haven't gotten used to it. I hope I don't. If you're in-house and feel like you stopped learning and can't name why, this is why. One company. One bet at a time. It was never your talent. It was how few bets you got to see. And the best part of all of this? I get to see these rascals a lot more than I used to.

Josh GrantSep 16, 2026

The biggest AEO mistake I’m seeing is publishing more content without understanding where recommendations actually come from. The full 2027 StackedGTM AEO playbook is coming soon. Here’s what I’d keep doing and stop funding, based on what I’m seeing actually working. START / KEEP DOING ✅ Follow the sources beyond your website. AirOps found roughly 85% of brand mentions came from third-party sources versus 13% from brand-owned domains in its commercial discovery research. I love this simple chart below. That isn’t a budget formula. Owned research can earn external coverage. But if most of your execution happens inside your CMS, investigate the mismatch. More content ≠ automatically better visibility. In more conversations than not, I’m seeing plans to publish more before anyone checks which sources actually appear in buyers’ answers. Find the external pages repeatedly cited when buyers evaluate your category. Identify where you have a credible case for inclusion and what evidence supports it. ✅ Understand whether you’re missing or being ruled out. If you’re absent, check content accessibility, relevant third-party coverage, and whether your positioning fits the buyer’s question. If you’re mentioned but rejected, read the explanation and check the cited sources. Is the objection accurate? An integration you shipped but never documented needs a public update. An integration you don’t have needs a product conversation. Treat the explanation as a lead to investigate, not proof of how the engine decided. ✅ Build from questions buyers actually ask. Pull recurring objections, comparisons, and requirements from sales calls, then use agents to surface relevant questions in public conversations on Reddit, X, and LinkedIn. Prioritize by frequency, buyer fit, and whether the question stalls a decision. You can build excellent Gong automations to flag those patterns and turn them into content briefs grounded in the customer’s own language. Publish specific answers with evidence, tradeoffs, and a clear explanation of when your product is a poor fit. STOP 🛑 Counting citations as recommendations. An answer can cite your research and recommend your competitor. Track citations and recommendations separately. Check whether you’re recommended for the buyers and use cases you care about, and how you’re described. 🛑 Reporting movement without checking what changed. Keep a consistent set of questions, repeat them, and separate results by buyer type and engine. Report new questions separately. Changing the questions can change your score. Answers also vary between runs. One favorable response after an update doesn’t prove it worked. 🛑 Accepting reports without follow-through. If an external comparison has your pricing wrong, assign someone to contact the publisher with evidence. If implementation details are missing, publish them. Record changes, recheck results, and track qualified demand alongside visibility. Dropping full 2027 guide soon