← All work

Case Study · 2021—present

Imersian — Spatial commerce for furniture

AI-powered 3D visualisation platform for furniture retailers. Co-founded and ran the whole company — product, engineering, marketing, sales, and operations — from 0 to live B2B partnerships.

Client

Imersian (Co-Founded)

Role

Co-Founder, CEO & CPO

Period

2021—present

Read

4 min read

Imersian — Spatial commerce for furniture

The bet

Furniture is the first physical commerce category that breaks at scale. You can't ship the showroom. Photos don't tell you whether the sofa fits. AR demos are ten years of hype and not yet a working product. Whoever solves the visualisation layer cleanly owns the rest of the buying journey.

We co-founded Imersian to take that bet. Bootstrapped. I own the cap table, not an advisory slot.

What we ship

Three products, one platform.

1. Imersian Visualiser (B2B, white-label)

The visualiser embedded into furniture retailers' own product pages. AI-generated 3D models from existing 2D product catalogues. Customers see the product in a generated room, swap textures, place it against their own dimensions.

2. Nestin + Imersian (D2C)

Consumer-facing interior design tool. Upload a room photo, get a designed scene populated with real, purchasable products. The AI plays art director; the catalogue does the conversion.

3. Imersian B2B Portal

Retailer onboarding, asset management, analytics. The plumbing that lets a furniture retailer onboard their entire catalogue in days, not quarters.

Architecture

  • Frontend: Next.js, React, TypeScript, Three.js, Redux
  • Backend: Node, MongoDB, AWS (SQS, Lambda, DynamoDB, CloudFront, S3)
  • AI pipeline: Spatial reconstruction layer + 3D-asset generation layer + LLM-driven catalogue extraction from Shopify product data — all wrapped behind a single internal API
  • Search & indexing: Algolia for product discovery, synced via automated pipeline
  • Distribution: White-label JS embed, native portal, public Nestin site, Shopify app

What running the whole company actually looks like

The product and engineering story is the easy part to describe. Here is what it looks like to run an early-stage bootstrapped startup with a 4–5 person distributed team:

Product & engineering — roadmap, UX research, prototyping, design, backlog, release planning, architecture decisions, code review, and direct contribution across frontend and backend.

Marketing & lead generation — full-funnel ownership: paid advertising (Meta, Google) from top-of-funnel awareness to bottom-of-funnel conversion. Apollo for outbound prospecting. HubSpot CRM for pipeline management, sequencing, and lifecycle automation. I built and ran all of it.

Sales — discovery calls, demos, proposals, negotiation, and close. Also post-close onboarding and ongoing customer success. No dedicated sales hire yet; I am the sales function.

Operations — financial modeling, P&L, unit economics, investor decks, due diligence materials, legal coordination, cap table management, team hiring across engineering, design, and operations.

Team management — hired, managed, and where necessary let go of people across all functions in a distributed team. Wrote the job descriptions, ran the interviews, set the culture.

This is not a CPO story with a CEO above it. There is no CEO above it.

Numbers (post-MVP launch)

  • +40% customer engagement vs. baseline 2D product pages (B2B partner data)
  • +50% growth in sign-ups quarter-on-quarter post-launch
  • 3 distinct products shipped from a single platform
  • Live B2B partnerships with established furniture retailers
  • Seed round in progress — pitch deck, financial model, and data room built

What this taught me about building something from nothing

Distribution is harder than product. The 3D visualisation technology is genuinely impressive. Getting a furniture retailer to change their product page workflow is harder. Marketing and sales have consumed as much of my bandwidth as engineering.

The model isn't the moat. The hard part is the asset pipeline — getting a retailer's 30,000-SKU catalogue into a state where the model can do something useful, without manual modelling at every step. That pipeline is the defensible IP.

AR is the wrong frame. Customers don't want to wave their phone at the room. They want to see the room finished. AR is a control mechanism; the product is the rendered outcome.

Dimensional fit converts; beautiful renders don't. "Looks good" drives engagement. "Fits my dimensions exactly" closes the sale.


Active: 2021 – present. Bootstrapped. Distributed team across Australia and Europe. Seed round in progress.

ProductSpatial AI0→1FounderCEO

Work with me

Building something like this?

If a productised version of this engagement fits your firm, the next step is a 30-minute discovery call.

contact@vencolab.com