Blog
Things we've shipped — and how.
Practical AI engineering. Architecture, guardrails, eval loops, and the production gotchas no one writes about. Filter by tag below.
- 11 min readstrategy · automation
Selling Digital Products With AI: From Course Creation to Delivery Without a Team
A solo creator with AI leverage can now run what used to take a five-person team: research, production, storefront, delivery, marketing, and support. But the same leverage flooded every marketplace with slop — so the money moved to whoever adds what AI can't. Here's the full stack, stage by stage, plus the quality gates that separate a product business from a refund queue.
Read → - 10 min readecommerce · content
Product Descriptions at Scale: 5,000 SKUs, One Brand Voice, Zero Copy-Paste Detection
47% of online sellers now use AI for product descriptions — and most of them produce the same interchangeable slop, or worse, still run manufacturer copy that five competitors also publish. Here's the pipeline that actually works at catalog scale: a codified voice, structured data in, template variation, per-SKU uniqueness from real attributes, and the QA gates that catch the AI's lies.
Read → - 11 min readagents · engineering
How We QA AI Systems Before Launch: Evals, Red-Teaming, and the Test Suite for Things That Talk
You can't unit-test a system that gives different answers to the same question. Here's the QA process we run on every chatbot, RAG system, and agent before it touches a customer: golden datasets, LLM judges, adversarial attacks, trajectory testing, and the release gate that decides ship or fix.
Read → - 10 min readaeo · strategy
GEO for Perplexity vs ChatGPT vs Gemini: One Page, Three Different Scoring Systems
Only ~11% of domains get cited by both ChatGPT and Perplexity. These engines don't share a playbook — they differ on retrieval architecture, freshness weighting, and source taste. Here's each scoring system decoded, and how to build one page that passes all three.
Read → - 10 min readaeo · strategy
GEO for YouTube: Why Video Is Now the Strongest AI-Citation Signal (and How to Exploit It)
YouTube is now the most-cited domain in Google AI Overviews — up 34% in six months — and brand mentions on YouTube predict AI visibility better than backlinks or domain authority. The catch: views and subscribers barely matter. Structure does. Here's the playbook for building videos like documentation and turning one recording into a citation machine.
Read → - 11 min readecommerce · agents
Fraud vs Friction: AI Return-Abuse Detection Without Punishing Honest Customers
Return abuse just became e-commerce fraud's #1 category — $76–103B annually, abusive returns up 64% in eighteen months, and half of consumers now using AI to write their refund claims. The panic response is tightening policy on everyone; the data says that backfires. Here's the risk-scored alternative: invisible for the honest 85%, escalating for the patterns — and the false-positive math that explains why kindness-by-default is the profitable setting.
Read → - 10 min readaeo · seo
Your About Page Is an Entity Signal: Brand Story Pages That Teach AI Who You Are
The About page has been marketing's junk drawer for twenty years — the founder photo, the 'our journey' timeline, the values nobody reads. Meanwhile, AI systems building answers about your brand treat it as a primary source: the page where you declare, in your own words and your own schema, who you are, what you do, and how to tell you apart from everyone with a similar name. Here's the rebuild — the entity layer, the disambiguation checklist, and the copy that serves humans and machines at once.
Read → - 10 min readecommerce · agents
AI in the Cart Drawer: Upsells, Shipping Thresholds, and the Highest-Intent Real Estate on Your Store
Every visitor who opens the cart is ready to buy — and most stores serve them a subtotal and a button. Benchmark data says AI-picked cart suggestions convert at more than double manual picks, slide carts drive a 3.2× revenue multiplier over cart pages, and the free-shipping bar is worth 10–20% AOV on its own. Here's the full playbook for the smallest, most valuable surface on your store — the mechanics, the AI layer, the mistakes, and the theme-level build notes.
Read → - 10 min readaeo · seo
Voice Search Didn't Die, It Moved Into AI Assistants: Optimizing for Spoken Queries in 2026
The 2018 voice-search hype died because assistants were dumb. Then the assistants got LLM brains: ChatGPT voice crossed 100M monthly users, Gemini replaced Google Assistant, Alexa+ went generative — and 8.4 billion voice devices now route queries through the same answer engines you're already optimizing for. Voice SEO and AEO just collapsed into one discipline. Here's what's genuinely different about spoken queries, and the checklist.
Read →
What we write about, and what we don't
Everything here comes out of engagement work rather than a content calendar, which shapes both the subjects and the tone.
- Numbers with baselines attached
- Where a post quotes a result, the before and after are both there. A percentage improvement with no starting point is unfalsifiable, and the AI industry currently runs on a great deal of it.
- The parts that were harder than expected
- Posts about builds include what went wrong — the retrieval that worked on curated documents and fell over on the real knowledge base, the eval case nobody thought to write. Those details are the reusable part; the happy path rarely is.
- Written for the person who has to fund it
- Most of these sit between a purely technical write-up and a business case, because that's where the decisions actually get made. Enough architecture to judge whether the approach is sound, enough cost and timeline to judge whether it's worth doing.
- No model-release commentary
- We don't publish reactions to every model launch. That writing ages badly and it's not what we know better than anyone else — what we know is what happens when these systems meet a real business process for six months.
