Why we exist.
Commerce has moved in one direction for twenty years. VISII exists because we spent fifteen of them in the one corner where that evolution breaks down.
Born where correlation fails.
VISII began in marketplaces, where the industry’s standard playbook breaks. When every piece is unique, purchase histories never repeat: “others also bought” has nothing to work with. Correlation, the engine behind most of e-commerce, fails exactly where choice is richest.
We could have moved to an easier corner. Instead the constraint became the company. To sell what has never been sold before, a system has to understand what a product means and what a shopper wants, from the first click. We did not choose understanding as a strategy; it was the only way through.
Fifteen years of R&D and a decade of client-by-client deployment produced everything we now sell: an intent model, patented in the US and EU, that turns a live session into a preference profile and knows browsing from buying; an architecture with no use for names or emails, built where there were none to collect; one engine rather than a stack of tools, because understanding compounds only when nothing strands it.
The discipline came with it: deployments are measured as randomised split tests against matched control traffic, read in the client’s own analytics. Nothing on this site asks to be believed; it asks to be checked.
VISII has been a team of engineers, researchers and commerce specialists attacking the same problem from different angles: what a product means, what a shopper wants, how a store should respond.
The angles now meet in one engine.
Principles.
Every major advance in commerce has asked less of the shopper. We believe the same transition is now reaching the operating side.
We build toward one shared understanding, not many isolated optimisations.
We measure everything against commercial outcomes, in the client’s own analytics.
We never ask anyone to buy promises instead of evidence.
The constraints we kept.
The engine took its shape from a handful of constraints, chosen early and kept since. Everything distinctive about VISII traces back to them.
The engine reads sessions, not people: what a shopper lingers on, skips and almost buys. It works without needing to know who anyone is.
The ten-day test is paid only if it clears the bar the client set. Our case studies and our commercial terms stand on the same measured number.
One engine, shaped by marketplaces where each piece is one of a kind, then deployed wherever taste leads the sale. The engine is the company.
One snippet to deploy, one snippet to remove. VISII earns its place surface by surface.
Our values.
If you’re ever thinking of joining us.
Arguments here end at the measurement, not at the loudest voice.
The marketplace taught us: master the hardest corner and the rest follows.
Prefer the change that makes every other part smarter.
The engine’s habit, kept by the people who build it: nothing learned stays private.
A signal captured anywhere should reach every surface that can use it.
Trust here is a number in the client’s analytics, renewed with every test.
Chosen by leaders in art, design, fashion and lifestyle e-commerce
As featured in
Frequently asked questions.
Do we really need an AI engine?
If your catalogue outgrew what your team can hand-tag and your shoppers can describe, yes. Most of the growth problems we are hired for, flat conversion, rising acquisition costs, merchandising that cannot keep up, are understanding problems wearing other names.
Is our catalogue too small, or too large?
The engine earns its keep from the high hundreds of items upwards and compounds as the catalogue grows. Smaller, taste-led catalogues still benefit, especially on mobile, where patience is shortest.
Which lever should we start with?
One engine, five commercial levers: average order value, conversion, engagement, traffic and costs. Start where your number hurts most; because each understanding teaches the others, the second lever is cheaper than the first. Most clients start with recommendations or search, and expand when it pays.
How do we integrate, and how long does it take?
Two routes: one lightweight snippet that sits on top of your existing site, or a RESTful API if you would rather build it in. Either way you are live within a single sprint, less than two weeks, on any platform; a product feed (Google Shopping included) and images are all we need to start.
What does a typical deployment look like?
Week one: we read your feed and images, and you see what the engine understood of your own catalogue before anything goes live. Week two: the first surface goes up behind a randomised split test, with success criteria you set in writing. From day ten you read the result in your own analytics and decide with evidence in hand.
How does the pricing work?
The ten-day test is £10K, paid only if it works. Beyond it, pricing is set to the size of the transformation, not to seats or searches: one predictable fee at full scale, and once the engine runs across the whole business, performance terms can enter the conversation.
Can it work with the platforms we already run?
Yes: Shopify, Magento, BigCommerce, Salesforce, WooCommerce or your custom build, headless or traditional. VISII works alongside your stack or replaces it surface by surface; nothing asks you to bet the stack on day one.
Can we test it against what we already have?
We insist on it. Deployments run as randomised split tests against matched controls; your analytics referee, not the incumbent and not us. If the test does not clear the bar you set, you pay nothing.
We sell in several languages.
The engine reads meaning from images and whatever text exists, in any language, with no dictionaries to maintain. Multi-language catalogues are normal deployments, not special projects.
How fast does it respond for shoppers?
Under 100 milliseconds in session. Understanding is only useful at the speed of browsing.
Any further questions?
Don’t hesitate to get in touch – a member of our team will be very happy to help.
Book Your Growth Call