01 — Embedded AI Operators
8-12 week sprints for ecommerce operators doing $3M–$10M. We embed, find where you're losing the most money, and ship working systems. Actual code. Not strategy decks.
Apply to work with usHow we start
Before any sprint, we do a proper diagnostic. We get into your ops, find where the waste is, and put a EUR number on it. You see the case before you commit to anything.
02 — Proof
Every system below was shipped while running live commercial operations. No sandbox. Real ops, real constraints, real before-and-after numbers.
40sec
Ads Reporting — 10 accounts, 6 currencies
Three people. Two days. One monthly report across 10 Google Ads accounts and 6 currencies. Built a Python pipeline with live FX normalisation. Same report, 40 seconds. Cost: $0.
48hrs
B2B CRM — EUR 18M division, 11 markets
EUR 18M B2B division, 11 markets, no CRM. Built a full pipeline and account management system over a weekend: deal tracking, multi-market data, the works. Salesforce's quote was $100K and 6 months.
1person
irth Studio — 274 products, 65 brands
274 products, 65 brands, pricing engine, SEO copy, product images, marketing strategy. Built solo. Pipelines handled data enrichment, image processing, title generation, and pricing. The kind of project that normally needs seven people.
9alerts
AI Support Agent — 4,342 SKU catalogue
4,342 SKUs, zero proactive monitoring. Built an AI agent that watches every support ticket, finds patterns by SKU and supplier, and fires 9 alert types: stock anomalies, fulfilment failures, quality issues. Problems caught before customers notice.
03 — What We Solve
We ran these operations ourselves. These aren't pain points from a survey. They're the five things that cost ecomm operators at your scale the most time and money, every single time.
Tool Sprawl
Prisync + Akeneo + Inventory Planner + Klaviyo + Nosto + your BI tool. Each one has its own login, its own data model, and none of them talk to each other. Your actual job ends up being copy-pasting between tabs. We build the connectors. No new tools required.
04 — How It Works
No retainers. No six-month "discover and align" phase that bills plenty and ships nothing. We diagnose before we build. Fixed scope, fixed price, working system at the end.
01
We get into your ops, find where the waste is biggest, and put a EUR number on it. You see the return before you commit to anything. We only take sprints where the numbers make sense for both sides.
02
We embed for 8–12 weeks. Find every manual process. Cost them out. Build the automations ranked by impact versus build time. You get a working system, not a roadmap. Shipped and tested, not planned and tabled.
03
Everything is documented and handed over. You own the code, the logic, the output. No SaaS dependency. No retainer. No check-in calls to keep the thing alive. We're done when it runs without us.
Investment
EUR 25,000 – 35,000
Fixed scope. Fixed price. No surprises.
Duration
8 – 12 weeks
Diagnostic to handover.
Availability
3 per quarter
Quality drops when you take on too much.
05 — The Team
Both of us have had real P&L on the line. Not advised on it. Run it. The systems we sell are the ones we built for ourselves, because we needed them and nobody else was going to build them.
Commercial AI Operator
Ran commercial operations at Desertcart: EUR 80M P&L, 60 markets, 15 currencies. Every time the instinct was to hire, the actual move was to build. 13+ AI systems later, those functions ran without a team behind them. CRM, P&L reporting, ads, support, email, catalogue. One person, right systems.
Commercial AI Operator
Bio and background to be added. Miquel's experience and key numbers will go here — commercial operations background, markets worked, systems shipped, measurable outcomes.
06 — Questions
Working code, documentation, and a handover session. The system runs without us from day one. No vendor lock-in, no monthly check-in, no dependency on us being around. Your team owns it and can maintain it without being technical.
A developer needs a spec. An agency hands you a deck and assigns it to a junior. We start from the business problem: what's costing the most, and why. Then we build the fix. We've run ops at this scale, so we don't need the problem explained to us.
$3M–$10M is the sweet spot. Big enough that bad ops cost real money. Small enough that you haven't got a full engineering team to fix it yourself. Above $10M? Still worth a conversation.
We start by understanding your ops: what tools you're running, where data lives, what your team does by hand and how often. We find the three biggest automation wins and put a EUR number on each. You see the business case before anything is signed. If the numbers don't work, we say so.
No. We connect to Shopify, WooCommerce, Google Ads, Klaviyo, whatever you're running. APIs handle the technical side and we handle the setup. We just need data access and someone on your side who knows the business problem. That's it.
Shopify, Shopify Plus, WooCommerce, Magento, BigCommerce. Google Ads, Meta, Klaviyo, Metabase. Anything with an API or a data export. Running something weird? Tell us in the form.
07 — Apply
Three clients per quarter is the limit. That's what proper attention allows. Tell us about your operation and where the waste is.
What happens next
You apply. We respond within 24 hours. If there's a fit, we book a diagnostic session, get into your ops, and show you the numbers. Then you decide.