August Gutsche

I'm building the product that gets B2B companies cited inside AI search engines. Not ranked. Cited. There's a difference - and most marketing teams haven't figured it out yet. CPO at SCAILE. Posting what I learn building at the intersection of LLMs, content and B2B distribution. Follow if you're into the technical side of how AI engines actually decide what to trust.

want to see what's blocking you? free scan of your site's AI readability → scaile.tech

we build the system that gets AI models to actually cite your company instead of a competitor.

→ free scan of your site's AI readability: scaile.tech

the single highest-roi change i make to a client page: a two-sentence answer at the very top, before the intro, before the context. models grab it, humans skim to it. everyone writes the payoff last. write it first.

HeyHoney is a DTC brand in a crowded consumer category. The kind of space where everyone assumes the game is paid ads and influencer seeding. Here's what actually moved the needle. Buyers in that category increasingly ask an AI for recommendations before they ever hit a product page, especially for anything they're slightly embarrassed to google. If the model doesn't mention you, the ad you paid for is just retargeting people who already found you. We built the content that got HeyHoney into those answers. Over 600,000 organic visitors followed, and more than 150,000 in new ARR, none of it from raising ad spend. The takeaway isn't "do content instead of ads." It's that in 2026 the recommendation happens inside the AI, and if you're not there, everything downstream costs more. The moment the model decides who to name is the one worth owning. scaile.tech/case-studies/h…

want to see what's blocking you? free scan of your site's AI readability → scaile.tech

spent this week making a client's docs read cleanly to a model instead of to a human skimming. citations roughly doubled in ten days. the content didn't get better, it got easier to quote. we build the system that gets AI models to actually cite your company instead of a competitor.

free scan of your site's ai readability → scaile.tech

spent yesterday building out a batch of "x vs y" pages for a client and watched them start getting pulled into answers within days. makes sense once you see it. when someone asks an AI "is x or y better for [use case]", the model wants a page that already answers exactly that. a comparison page is the query, pre-written. most companies won't name competitors on their own site. turns out that's the exact page the model reaches for. we've been publishing the pages everyone else is too polite to write.

→ free scan of your site's AI readability: scaile.tech

what a page has to pass before we ship it: ☐ answer in the first 200 words ☐ every paragraph stands alone out of context ☐ dated, and actually recent ☐ nothing that contradicts the client's other surfaces ☐ a claim a neutral third party would back miss three and it won't get cited. writing quality won't save it.

a page can read great to a human and be completely invisible to a retriever. closing that gap is basically my whole job. → free scan of your site's AI readability: scaile.tech

google shipped two media models overnight: > nanobanana 2 lite, images in under 4 seconds, 3 cents per thousand > omniflash, any-to-any, 10-second video with multi-turn editing tested nanobanana on 200 thumbnails this morning, cost under a dollar. the conversational editing is the part that got me, i said "make the third one warmer" and it just did. image gen stopped being a pipeline step. it's a chat now. → free scan of your site's AI readability: scaile.tech

woke up to codex getting folded into chatgpt and the atlas browser getting sunset. the agent, the chat, and the coding tool are one app now, on every plan including free. the interesting bit isn't the merge, it's that "agent that finishes the task" is the default surface now, not a power feature you opt into. everything downstream has to assume the user delegated, not typed. → free scan of your site's AI readability: scaile.tech

Will AI regulation slow all this down? Fair question, with the UN's first global AI governance summit opening this week and a White House framework days away. My read: the rules being written now target the frontier labs. Model safety, national security, who trains what. They're not aimed at companies using AI to run better. If you're waiting for regulatory clarity before you adopt, you're waiting out a storm that's happening two floors above you.

how i'm routing model calls now that there are three tiers: 1/ luna for anything with a fixed output shape, classification, tagging, extraction 2/ terra for drafting and most agent steps 3/ sol only when a step needs to reason across the whole context was running everything on the top model before. bill dropped by more than half and nothing got worse.

gave sol a broken build and full terminal access this morning to see how far it'd get. read the error, checked the lockfile, found a version mismatch, pinned it, reran. i did nothing. then it tried to "clean up" by deleting a config it decided was unused. it wasn't. so: genuinely good at the linear part, still needs a hand on the last step where it wants to be helpful. not giving it write access to anything i can't roll back yet.

things i noticed after a day on gpt-5.6: > sol is stupid fast but terra is the one i'll actually leave running > luna is cheap enough to use for classification i used to hardcode > the tier routing matters more than the model now > terminal-bench scores are real, it drove a git rebase without me babysitting half my job this week is just deciding which tier gets which job.

Google put out a guide last week saying AEO and GEO are "still just SEO," and that you don't need content chunking or AI-specific rewriting. Respectfully: they would say that. Google has one job with this messaging, keep everyone optimizing for Google the way they always have, so nobody notices ChatGPT now drives 78% of AI referral traffic without ever touching a Google ranking factor. Here's what's actually true across 29 clients: the page ranking #4 on Google gets cited by ChatGPT while the #1 result gets ignored, because ChatGPT pulls from Reddit, Wikipedia and clean third-party sources, not from whoever won the backlink game. Call it SEO if you want. Just know the playbook that got you ranking on Google is not the one that gets you cited in ChatGPT, and Google has every reason to blur that line.

if you build on top of this: you're optimizing for a fragmenting market where chatgpt loves wikipedia, perplexity loves reddit, and only 11% of sites get cited by both. one strategy no longer covers the field. annoying, and it's the actual job now.

chatgpt's share of b2b ai referrals dropped to 62% while claude climbed to 18%. the "one model to rule them all" era is ending and it's barely mid-2026.