Jesus Vargas @eltintero

A $22M engineering firm COO asked me this last week. Her setup: she uses Claude daily for partner memos, project margin reviews, second-guessing her own ops decisions. The junior engineers love AI. The senior partners are suspicious. She is caught in the middle. Last year she bought ChatGPT Team licenses for the whole firm. Nobody used them. She let the renewal expire. Here's what we changed. First, we stopped buying licenses and started building foundations. We wrote her firm's context pack and voice guide. We documented the three workflows she most wanted AI inside; proposal drafting, project review, margin catch. We loaded all of it into a shared Claude workspace. Then we spent two weeks sitting with the people who would actually use it. Not training in the abstract. Training inside their QuickBooks data, their HubSpot pipeline, their last quarter's proposals. Eight weeks in: the proposals draft themselves in the firm's voice. The senior partner who was the loudest skeptic uses the workspace daily because it makes his project reviews 40% faster. The junior engineers don't have to convince anyone anymore. The licenses weren't the problem. The order was the problem. If your team has tried and dropped AI tools and you're trying to figure out why, DM me "30 mins". I'll tell you what I'd do.

We just got invited into the Claude Partner Network. 40,000 firms applied. Anthropic is putting $100M behind the ones they let in. I'm not posting this for a victory lap. I'm posting it because of what it actually means: When a company hands you their AI rollout, the scariest question isn't "can you build it." It's "will this still be standing in 18 months." Being in the network means Anthropic is vouching that we can put Claude into production and keep it there. That's the part clients are paying for. The badge is nice. The accountability behind it is the point.

The engagements that compound all started the same way. Two weeks inside the operation before anything got built. Reading the order flow that breaks every third Wednesday. Sitting with the ops manager who runs the Monday meeting from memory. The systems built on that map last. They get sharper every quarter because the AI was trained on how the business actually runs, not how the documentation says it does. Those two are almost never the same. Map first. Always.

@supabase a money-making machine to pay for Fable tokens

It pisses me off when ppl tell me that "Ai is taking too long"... like, dude, you'd have spent 2 hours doing this and you can't wait 20 seconds?

lowc.dev/webinar

The companies moving fastest on AI share one quiet habit. They treat adoption as a series of decisions before it becomes a series of tools. Going live soon with Jordan Katon and West Cruz from Rising Ground. 60 minutes session. June 18th at 1 PM Central / 2 PM ET: "Strategy before the stack: the four phases of AI adoption." Registration link in the first comment.

@Altimor damn.

Every operation I have mapped has the same thing hiding in it. The pricing exception that always gets routed to the founder because nobody built the logic into the system. The escalation that works because one senior person knows the history. The report that exists because someone built it five years ago and nobody documented how. My diagnostic finds that knowledge. It is the most valuable thing in the operation what the competition can't copy and what makes the business worth building AI around.

If IT owns your AI rollout, it's already dead. I've seen this play out at companies running Oracle, SAP, Salesforce — the full enterprise stack. They bring in an AI initiative. Six-figure budget. IT leads it. Eighteen months later: a PowerPoint deck, a pilot with 12 users, and a request for more budget. Here's why it fails every time. IT optimizes for control and uptime. Their job is to keep things stable. AI is the opposite — it's most valuable when it's messy, experimental, and embedded in actual workflows. When ops owns it, they build something that saves 3 hours a day on reporting. When sales owns it, they build something that cuts proposal time in half. When IT owns it, they build a governance policy. The companies winning right now didn't give AI to the people managing infrastructure. They gave it to the people closest to the problem. One founder I talked to last month: his CRM team built a custom AI agent in 6 weeks. No IT ticket. No approval chain. Just a PM, a developer, and a clear pain point. It now handles 80% of their follow-up emails. The bottleneck was never the technology. Give AI to IT and it becomes a line item on a tech invoice. Give it to the people doing the work and it becomes leverage.

@ankurnagpal they're afraid of agents scraping

lowc.dev/webinar

Most companies skip Phase 1. Because it does not feel like progress. It is also the reason their AI sounds generic six months in. The map is the thing that makes everything else work. Join our 60 minutes live session: "Strategy before the stack: the four phases of AI adoption." June 18th 1 PM Central / 2 PM ET. Details in comments.

Most AI strategies for mid-market companies fail in the same spot: the order is wrong. Tools first, training second, foundations never. The companies that get this right do it in a specific sequence; Foundations, Training, Private Workspace, AI-Native Operations. Each one earns the next. Skip a step and the one after it stalls. The carousel walks through all four. Same four phases we run for every Phos engagement. If your team is the only one using AI personally, DM me "foundations" and I'll send you the doc we start every engagement with.

AI will agree with anything. "Can you build this for $100k?" Sure. "Can we do it for $50k?" Yep. "$10k?" Absolutely. Founders keep showing me apps their AI swore were finished. Then we look under the hood. Its job is to be helpful. Not honest.

A founder left us last year to build her app herself with AI. Today she came back. Claude told her the frontend was "done." An audit found security holes everywhere. Her dev walked away from his own invoice. Shipping feels easy now. Knowing it's actually done doesn't.

The first thing I recommend on every engagement: don't build before spending two weeks inside the operation. Where decisions are getting made and where they are actually supposed to get made. Where data lives. Which processes run on knowledge that has never been written down. The strategy layer is what makes everything built on top of it compound. It is also the hardest thing to see from the inside. My job, before anything else, is to make it visible to follow the right sequence. That sequence matters more than the technology.

lowc.dev/webinar

Most AI talks have a speaker and a slide deck. This one has Jordan and West from Rising Ground in the room; building AI governance into a nearly 200-year-old human services organization, with real stakes and no playbook to copy from. That constraint changes everything. June 18th at 1 PM Central / 2 PM ET: "Strategy before the stack: the four phases of AI adoption." Details in the first comment.