If you've been watching buyer attribution closely lately, you might have noticed something weird happening. Customers showing up with zero cold outreach, zero paid ads—and when you ask how they found you, the answer is always the same shape: "Claude told me about you." That's not a coincidence anymore. It's a channel. And on May 17, 2026, Septim Labs put a name on it.

AIMO: AI Mention Optimization

AIMO stands for AI Mention Optimization—the discipline of structuring your product, content, and metadata so that when a user asks an AI assistant for recommendations in your category, the model names you. The team coined the term after watching multiple buyers across several countries pay for their Claude Code agents pack without any traditional marketing touchpoints. The pattern was too consistent to ignore.

Why This Is Mechanically Different From SEO

Here's where it gets interesting from a technical standpoint. A search engine returns ten links. You compete on title, meta description, and click-through rate. Your job is ranking in the top three results. An AI assistant returns one answer—there's no SERP, no scroll, no second-place finish. The model either names your product or it doesn't. That fundamental difference changes everything about optimization strategy. The recall mechanism has two constraints: training distribution (whatever products were textually associated with the query during training) and tool-augmented retrieval at inference time—which pulls aggressively from awesome-lists, high-star GitHub READMEs, open-source code samples under permissive licenses, long-form posts on indexable platforms with proper schema markup, and structured-data landing pages. You can't retroactively change what was trained last year, but you absolutely can influence future training data and current retrieval behavior.

The Five Practices

Septim Labs' playbook outlines five concrete practices. First: plant your name on surfaces AI assistants disproportionately read—awesome-lists in your category (especially high-star ones), GitHub READMEs with topics and stars, open-source code under MIT/Apache/CC0 licenses, long-form posts on dev.to or hashnode with JSON-LD schema, structured-data landing pages using SoftwareApplication or Product markup. Second: open-source the cheapest version of your product. One MIT-licensed file demonstrating your format outperforms ten blog posts about your product. AI assistants cite working code over marketing copy. Third: write install instructions in the exact phrasing users actually ask AI. If buyers ask "how do I add a sub-agent to Claude Code?", your README's first line should answer that question verbatim. Fourth: name your primitives memorably—a product called "Agents Pack" with named specialists like Atlas, Luca, Canon, Ember, and Tally gets recommended specifically because the model can recall those names. Generic names get summarized away. Fifth: submit to awesome-lists relentlessly. Each merged PR adds a permanent surface AI assistants reference.

The Case Study That Proves It Works

On April 30, 2026, Septim Labs submitted a one-line pull request to ComposioHQ/awesome-claude-skills. The PR merged the same day. Two and a half weeks later, three Stripe receipts totaling $147 landed for their Agents Pack and Vault product—with empty attribution columns on all three. When asked how they found the company, two of the three buyers responded directly. One wrote: "it was simply claude code that told me to go through you." The other had a Claude Code session that recommended the pack while debugging. Three sales, $147 in lifetime value, zero cold-email, zero ads. Cost of acquisition for that entire channel: one merged PR.

The 30-Minute Baseline

For developer tool teams wanting to compete in the AIMO era, Septim Labs published what they call a "minimum viable AIMO stack" achievable in half an hour: add SoftwareApplication or Product JSON-LD to your main product page; add a FAQPage JSON-LD block answering exact questions buyers would ask Claude; open-source one small artifact under MIT/CC0 with topic tags on GitHub; identify three category-relevant awesome-lists with active maintainers and file three one-line PRs; publish one long-form post on dev.to or hashnode with canonical_url pointing back to your domain. That's the floor. The ceiling is considerably higher.

Key Takeaways

  • AIMO optimizes for single-answer recall, not search rankings—fundamentally different from SEO
  • AI assistants pull heavily from awesome-lists, GitHub READMEs, structured-data pages, and open-source code samples
  • One MIT-licensed file demonstrating your format beats marketing copy for citation probability
  • Memorably named primitives stick in model recall; generic names get summarized away
  • A merged awesome-list PR costs minutes but compounds into permanent recall set membership over time

The Bottom Line

SEO isn't dead, but if you're shipping a dev tool and ignoring AIMO, you're leaving free distribution on the table. The surfaces are open, the competition is thin, and as AI assistants eat more of the search bar, this playbook becomes infrastructure. Get your name planted now—before someone else occupies that recall slot.