Oak Grove
Solo operator · US, UK & Japan · ~10 AI-assisted workflows
I started Oak Grove in 2022 to test a premise: could one person build and run a home-decor brand almost entirely solo, using AI across the whole operation? Almost every function is handled by one person, me: sourcing, design, photography, copywriting, translation, publishing, marketing, and analytics.
Products are made on demand and shipped by fulfillment partners. This case study covers how I used AI, and how that use changed as the tools evolved.
Topics
Oak Grove · Shopify storefront · US market
The brand
Authentic Japanese stencils, built into a home goods brand
Oak Grove transforms authentic 19th-century Japanese stencils into woven jacquard blankets, throw pillows, and other home decor goods. Each pattern begins as katagami: a hand-carved stencil used for dyeing traditional textiles like kimono, furoshiki, and nōren.
After launching in the US, the brand expanded to the UK (2024) and Japan (2025). Everything was built from scratch: brand identity, full product line, multi-market Shopify infrastructure with English and Japanese storefronts, and a sourcing and curation pipeline for stencil artwork.
Original katagami stencil
Woven blanket
Throw pillow
Decorative tray
2022 – 2025
AI via the chat interface
From 2022 through 2025, I used AI entirely through chat interfaces — ChatGPT, Claude, and Gemini — for research, copywriting, translation drafts, and market analysis. No pipelines, no API calls.
| What Worked Well | What Didn't Work |
|---|---|
| Product descriptions, ad copy, translation, and data analysis | Paid ads — AI guidance on campaign structure was often confidently wrong; primary source documentation worked better |
| Generative image tools (eventually) — held off on early outputs due to uncanny quality; adopted once models improved enough that images looked like what a customer would actually receive |
Four years of learning which tasks benefited from AI and which required human judgment shaped every design decision in the 2026 workflow layer.
↗ Me, Myself, and AI: 3 Years Building an E-Commerce Brand Solo — LinkedIn article (published March 2026)
2026
AI-assisted workflows built in n8n
In 2026, I replaced ad hoc AI prompting with structured pipelines built in n8n. Most trigger from Airtable, call AI models via API, and write results back for human review before anything is published or sent.
Workflow 01
Inspiration Stager
Before
No equivalent process. Writing staging prompts meant starting from scratch each time, describing lighting, furniture, geometry, and mood in text, with no guarantee the output matched what I'd imagined.
After
I submit a room photo I like alongside my product photo. The inspiration image replaces the prompt: Gemini reads the scene and places the product into it. The stencil pattern and colorway are preserved exactly as they appear in my photo.
Inspiration Stager Workflow
Airtable record: inspiration image, product photo, AI-generated scene prompt, and final staged output
Workflow 02
Lifestyle Mockup Generator
Before
Generated lifestyle images one at a time in the Gemini chat UI, manually writing a prompt for each room setting, then repeating. Time-consuming and inconsistent across products.
After
One trigger run produces four room variants from a single sample photo. Room-specific prompts are constructed automatically per product type. All outputs land in Airtable and Dropbox for review before use in listings or ads.
Mockup Generator Workflow
Input: sample photo I took
Output: mockups for each of 4 scene variations
Workflow 03
Colorway Pipeline: Three Linked Workflows
Captures color inspiration from anywhere, applies it to a stencil to generate up to 6 colorway variants, then routes the results through a curator and a reviewer before anything moves to production.
Before
Chose colors by instinct, like using a color picker as a Ouija board, guided by gut feeling with no record of what I'd tried and no way to compare options. I'd create products using whatever colors I thought looked good and upload them directly to Shopify.
After
I save a photo of a pillow I like the color scheme of from anywhere: a browser on desktop or my phone. It lands in Airtable automatically with the extracted hex codes. I use those palettes to generate pillow mockups in up to 6 colorways, then pass the promising ones to reviewers for feedback before anything goes to production.
Step 1 of 3 · Palette Lab
Capture color inspiration from anywhere
Saves color inspiration from photos into a reusable palette record. Photos submitted from desktop or mobile are analyzed by Gemini, which extracts the hex codes and writes them to Airtable, ready to be used by the Colorway Generator.
Palette capture workflow
Grid view
Gallery view
Step 2 of 3 · Colorway Generator
Apply a palette to a stencil and generate up to 6 variants
Takes a validated palette from Palette Lab and a black-and-white stencil design. Gemini applies the palette to the stencil and renders up to 6 colorway variants on a product mockup, fast enough to judge which combinations work before committing. It's a preview tool, not a production one: Gemini approximates the recolor well enough to evaluate, while exact, production-accurate color still happens in Photoshop.
Colorway generation workflow
Step 3 of 3 · Design Review
Curate a batch, send to reviewers, collect yes / maybe / no votes
Two small web apps I built with Claude. On the first, I review the finished colorways and pick which ones are worth showing. The second goes to family reviewers, who vote Yes / Maybe / No on each design and can leave a note. Every vote is saved automatically, so I can see what resonated before committing anything to production.
Feedback loop workflow
Curator view: select colorways to send for review
Reviewer view: Yes / Maybe / No per design
Other Workflows
Additional automation across content, publishing, and operations.
What this demonstrates
Strategic judgment, human-AI systems, built on four years of solo operation
Those gates came from specific failures, not theory. The Inspiration Stager runs two sequential Gemini calls (the first reads the room and is explicitly told to ignore the product, the second places the product into the scene) because a single combined call reimagined the stencil to fit the room. The Colorway Generator stays a preview tool for the same reason: Gemini approximates a recolor well enough to judge a combination, but exact color still happens in Photoshop. Each design decision traces back to a model failure that mattered.
The judgment behind these decisions came from four years of running the brand solo and using AI through chat interfaces, long before any of it was automated. The workflow layer itself is new: built in 2026, revised when it produced bad outputs, and rebuilt as better models shipped. That pairing is the point: durable judgment about where humans belong in the loop, applied to implementations kept deliberately current.