Technology & Innovation

CTO Checklist: Outsource Room Preview Versus Building In-House?

9 min readAI FurnitureShopify furniture widget

The AI Furniture room-preview widget is what we build, and live Shopify demos such as Mirage Furniture already use it. What we actually ship is what these articles document: photo upload on the PDP, composites in ~30 seconds from catalog JPEGs you already have, Shopify app install or jsDelivr, and 10p per billable preview. This is not marketing theory copied from other blogs.

Room preview build versus buy as a CTO decision

Placement + lighting + Shopify theme UX + billing + abuse controls, once built, form a product rather than a weekend notebook demo.

When buying makes sense

  • Preview on PDPs is needed this month
  • Dedicated ML + frontend capacity is lacking
  • Predictable per-preview COGS is the goal (our public plan is 10p)

When building makes sense

  • Models, evals, and on-call will be staffed
  • Exotic catalog constraints that a vendor cannot meet are required
  • Delayed revenue is accepted in exchange for control

Buy, instrument, then revisit build in 12–18 months with real usage data — that path fits most mid-market furniture CTOs.

Frequently asked questions

What justifies building in-house?

Building in-house is justified when visualization counts as core IP, ML/platform staff already exist, and multi-quarter time-to-PDP is acceptable. Buying first is what most furniture merchants should do.

See it in your room

Install costs £0. Signup includes 50 free room previews. After that, pick Starter (£39 / 80 included), Growth (£69 / 250), or Pro (£99 / 500) — then 10p per extra preview. Enterprise is custom pricing for large catalogs. Existing stores that already have 500 free previews keep that pool.

CTO Checklist: Outsource Room Preview Versus Building In-House? | AI Furniture