Built a luxury product brand from concept to launch in 26 days using AI

Project Description

Ilmur ("Scent of Iceland") is a luxury souvenir brand launching hand-cast ceramic diffuser kits inspired by Icelandic basalt columns (Stuðlaberg). Each kit contains a porous clay diffuser with internal essential oil pipes, two meditation clays, and a 5ml bottle of organic Icelandic essential oil from supplier Hraundís — representing four iconic locations: Þingvellir, Jökull, Eldfjall, and Norðurljós. Every SKU is limited to 100 hand-numbered units with a unique clay form variation, sold in a charcoal rigid box with navy velvet insert, priced at 89,000–149,000 ISK ($640–$1,070 USD). The product ships dry — customers apply oil to the porous clay, which diffuses scent for 1–3 days. A special 5th SKU, "Pink Stuðlaberg," adds a premium handwritten card and unique pink clay formulation. Presale launches April 2026, production July–August, delivery October 2026. The entire project — from concept to launch-ready infrastructure — was built in 26 days by one person (with a design partner) using AI as the complete product development, marketing, finance, and operations team.

Achievements

In 26 days (January 20 – February 15, 2026), I created the complete infrastructure for a luxury product launch using AI:

- Investor-grade project plan (v2, 20+ pages): WBS, 32-week Gantt, 8 phase gates, RACI matrix, 12-risk register, unit economics, launch plan

- 3-scenario financial model: break-even at 20 units, 62–78% gross margin, NRE budget 560k–1.8M ISK

- Competitive pricing benchmarking against Diptyque, Trudon, L'Objet, and Fischersund

- 3 distinct sales pitch variations (sensory, mythic, direct luxury positioning)

- React presale landing page with live inventory counters and Shopify integration (LIVE)

- 20+ AI-generated product images: hero shots, 9-angle packshot grids, lifestyle scenes, bottle close-ups — all before a physical prototype exists

- Cinematic product film script with Canon R5 specs and shot-by-shot direction

- Complete Higgsfield production workflow guide (tool selection, camera presets, prompt templates)

- 3 campaign concepts: "You Are Holding Iceland," "The Gateway," "100 Fragments"

- Full brand narrative weaving Jules Verne, Snæfellsjökull mythology, and Icelandic folklore

- Supplier research dossier on Hraundís with 4 scent formulations, NDA draft, and outreach email

- EU CLP compliance workstream: labeling requirements, SDS structure, allergen declarations

- Kraum retail pitch deck outline with margin analysis

- NotebookLM podcast-style audio overview and infographic

- Confirmed design partner (Emilia Telese, internationally known artist) and identified manufacturing pathway

Tools

- Claude (Anthropic) — Primary strategic partner: project plan, financial model, risk register, RACI, sales pitches, presale page code, brand narrative, campaign concepts, compliance workstream, supplier outreach, retail pitch, submission letter

- Higgsfield — Product photography and video: Nano Banana Pro for 4K packshots, Angles 2.0 for multi-angle consistency, Edit Image for label corrections, Cinema Studio v1.5 + Camera Controls for cinematic video

- Google Gemini — Concept art and product renders: Pink Stuðlaberg hero shots, Icelandic landscape compositions, atmospheric lifestyle scenes

- Google NotebookLM — Audio overview (podcast-style project deep dive), infographic, visual mind map of project structure

- Lovable — Web development: presale landing page with React components, live inventory counters, Shopify integration

- Shopify — E-commerce platform for payment processing and order management

- CamScanner — Digitization of original 7-page handwritten concept notes

Methodology

I applied 40 years of infrastructure project management discipline to a luxury product launch — using AI as my entire team.

1. Origin capture: Wrote 7 pages of handwritten concept notes (blue ink, kitchen table), then scanned and used them as the project's "source of truth" for every AI interaction.

2. Strategic foundation first: Before any visual work, I used Claude to build the complete business architecture — unit economics, pricing benchmarks, risk register, compliance requirements, supplier analysis. This gave me the discipline to know what "good" looks like before generating creative assets.

3. Iterative AI partnership: I don't use AI as a one-shot generator. I push back, challenge assumptions, and iterate. When Claude's first project plan had timeline contradictions and missing compliance workstreams, I had it audit itself against the original brief and rebuild. The v2 plan is the result of that critical feedback loop.

4. Visual prototyping before physical production: Used Higgsfield and Gemini to create the entire visual identity of the brand — 20+ images, packaging concepts, bottle designs — before spending a single króna on materials. This de-risks the product by testing positioning and aesthetics with potential buyers before committing to manufacturing.

5. Presale-funded model: The entire business model eliminates upfront capital risk. The presale campaign (April 2026) funds production. If 60 units/SKU don't sell, production doesn't start. AI made it possible to build all the marketing infrastructure needed to run this presale without a team or agency.

6. Phase-gated execution: 8 Go/No-Go gates across 32 weeks, each with explicit pass/fail criteria and contingency plans. Engineering discipline applied to a creative product — because beautiful things still need to ship on time and on budget.

Learnings

- AI does not replace taste. Every product decision — the basalt column form, the Snæfellsnes clay, the Jules Verne storytelling angle, the numbered limited edition model — came from human instinct, cultural knowledge, and 66 years of living in Iceland. AI amplified the execution. The soul of the product came from the soul of its maker.

- AI collapses the team problem. A luxury product launch normally requires a product designer, ceramic engineer, scent formulator, packaging designer, compliance specialist, photographer, copywriter, web developer, financial analyst, and project manager. AI gave me all of them in one window. Not perfectly — but well enough to build the entire launch infrastructure before production begins.

- AI rewards experience. My engineering degree, MBA, and four decades managing complex international projects made me a dramatically better AI user. I knew what a real project plan looks like. I knew when a financial model had weak assumptions. I knew when a risk register was missing a scenario. Experience is the multiplier that turns AI from a novelty into a professional tool.

- Push back on AI or it will give you mediocrity. The first project plan Claude produced was 80% solid and 20% contradictory. I caught the contradictions because I've reviewed hundreds of project plans in my career. AI gets better when you treat it like a talented junior associate who needs senior review — not an oracle.

- Visualize before you build. Generating 20+ product images before a prototype exists is the single highest-ROI use of AI in this project. It made the product real for my design partner, for potential retail buyers, and for my own decision-making. You can't sell a feeling with a spreadsheet. You need an image.

Relevant URLs

https://ilmur-iceland-scent.lovable.app