Skip to main content

Systems in production.

Two systems AILoopwise built that run today: what they do, how they are built, and the numbers straight from production.

Fassadenklar: from a CRM deal to a priced quote, with the building measured from public data

Customer: Fassadenklar (fassadenklar.de) — facade cleaning and coating, Germany. What we built: Pipedrive deal → automatic building measurement from official 3D city data → draft quote in sevDesk. Status: the measurement pipeline has been in production since April 2026. Claude has been running on Fassadenklar's own Anthropic account since September 2026.

The situation

Quoting facade work means knowing the surface area of a building. It used to mean driving there, measuring on a map, or guessing — for every enquiry. The slow step was not the work, it was the paperwork before the work.

The quoting software originally planned never opened its API. We did not wait; we built the pipeline on sevDesk instead.

What we built

  1. A deal is marked for a quote in Pipedrive. A webhook fires.
  2. The address is located via the Google Geocoding API, and in Berlin via the official address register.
  3. The building is measured from official LoD2 city models — the same 3D data the surveying offices publish. No site visit, no manual tracing.
  4. A vision model reads the images: Google Street View, the customer's photos from the deal, and a Google map with the building outlines drawn in. From these it derives storeys, balconies, surface material and access.
  5. Two models check each other. Where the match is ambiguous, Gemini and Claude identify the building independently on the map. Only what both agree on is used.
  6. The draft quote is created in sevDesk, positions and prices from Fassadenklar's fixed template. No model writes a price. A quote is a contract, so the contractual parts are deterministic by design.
  7. The team reviews and sends. Sign-off stays with Fassadenklar.

The pipeline in numbers

  • Over 50,000 m² of facade measured automatically. Since September, each measurement becomes a draft quote in sevDesk directly.
  • 14 of 16 German federal states connected, around 94% of Germany's population.
  • 0.2% deviation from the on-site survey: three apartment buildings in Dortmund, 2,088.9 m² measured against 2,093 m² surveyed by hand.
  • Over 80 production releases since April, backed by over 600 automated tests.

From operation

Two examples:

  • Leonberg — measured: two buildings, 245.18 m² and 187.36 m². One quote covering both: €3,460.32 net.
  • Berlin — measured: 849.42 m² of facade, with a 112.35 m² party wall automatically excluded from billing. Quote: €6,795.36. A party wall sits on the boundary and cannot be charged. The system catches it automatically — at 8pm too.

The rule decides, not the model

Claude writes a paragraph about the specific building for every quote. A rule guard ensures that areas and prices come only from the measurement and the template, never from the model. Every quote pairs a personal text with exact figures.

Book your Growth Assessment

Our own company

Live in the Movent: loyalty, gamification and ticketing, built with Claude Code

Our own platform, not a client: Live in the Movent (liveinthemovent.com) — a loyalty and gamification platform for live events: ticketing, a digital ticket wallet, collectible 'Movents' for every event a fan attends, and points and rewards. For independent organisers and nightlife venues, Frankfurt. Operated by Unique Moments UG (haftungsbeschränkt), Stuttgart. What we built: the backend, the fan web app and iOS app, a scanner app for door staff, an organiser portal, and the marketing system that researches and ships our own SEO content and social posts. Status: in production. The app has been live in the App Store since June 2026.

The situation

We wanted a real product to prove the method on, not a slide deck. Ticketing for independent nightlife organisers is a genuine problem: small venues get little control over pricing and fees, no loyalty data, and no direct relationship with the fans who come back every week. So we built the whole thing, backend to iOS app, together with our advisors Yu-kai Chou, Cornelius Ballin and Benedict Windelband; the gamification design follows Yu-kai Chou's Octalysis framework. We ran the marketing the same way we build the product: Claude does the work under our review, not as a chat window on the side.

What we built

  • Backend — the API behind events, ticketing and payments (Stripe Connect, PayPal), with multi-tenant accounts for each organiser.
  • Fan web app — the marketing site, ticketshop, dashboard and blog.
  • iOS app — the fan-facing app: the digital ticket wallet, a collectible Movent for every event a fan attends, and points and rewards. Live in the App Store since June 2026; an Android version is in development.
  • Scanner app — ticket and reward scanning for event staff at the door.
  • Organiser portal — event, ticket and coupon management for the venues and promoters who run events on the platform; a rebuilt version of the portal has been Claude-co-authored from its very first commit.
  • Marketing system — a separate Claude Code system that researches, drafts and publishes our SEO blog content and social posts. It has produced 35 published SEO articles on the site, and 97% of its own commits are Claude-co-authored.

Across the six repositories behind all of this: over 6,800 commits, 44% of them co-written with Claude.

Where it stands

Pulled from the production database, as of September 2026:

  • Over 8,000 tickets sold, for over €240,000 in ticket sales.
  • Over 3,800 people have bought a ticket.
  • 12 organisers are actively running events on the platform.

What's next

An Android release for the fan app.

Book your Growth Assessment

No company should leave its growth untapped.

See how much potential is in your organization.

We only take on companies where we see clear growth potential. In a 30-minute assessment we find out whether that applies to you. If it doesn't, we'll tell you.