Iris.
A custom operating system for how your business actually runs. Liquid software whose every part is shaped to you.
One foundation, shaped to you.
Most businesses run on a dozen disconnected tools that never quite fit and never quite talk. Iris is the opposite: a single custom operating system, built around how your business actually works, with the parts you need and none of the ones you don’t.
We call it liquid software because nothing is fixed. CRM, knowledge base, sales, ERP, projects, the connective tissue between them, every part is customisable and reshaped to your operation rather than bent to fit someone else’s template.
core modules in one system
CRM · KB · Sales · ERP · Projects · yours
isolated infrastructure per customer
Per-customer, governed
off-the-shelf templates
Every part shaped to you
Liquid software whose every part is shaped to you.
Customisable parts, one system.
CRM
Contacts, pipeline, and the relationship history your team already keeps in five places.
Knowledge base
The shared brain: docs, processes, and answers, retrievable by people and by AI.
Sales
Quotes, deals, and follow-ups wired to the rest of the business, not a separate silo.
ERP & operations
Orders, resourcing, and the day-to-day mechanics, modelled to your actual process.
Projects
Work, owners, and timelines that read from the same data as everything else.
Your modules
Whatever your business runs on. Every part is shaped to you, not forced into a template.
Scattered data, made retrievable.
Underneath the modules, Iris runs a retrieval database, the layer that decides whether AI on your business is useful or a party trick. It centralises what you already have, cleans it, and structures it so an agent retrieves the right record with the right permissions rather than guessing at keywords across a pile of files.
Centralise
Pull data out of the dozen systems it's scattered across into one retrieval layer, without forcing a migration first.
Clean
Resolve duplicates, fix the mess, and reconcile the contradictions, the work that quietly breaks every AI project.
Categorise
Structure and tag everything so retrieval is accurate and permissioned, not a keyword guess across a pile of files.
No forced migration to begin. The retrieval layer reads what you already run; replacing a source system is a decision you make later, on its own merits.
Your data, ready for AI.
Because everything lives in one system, your data is finally in one shape, which is exactly what AI needs. Iris is built around a clean data core, so agents can read across your whole business and act inside it, with the right permissions and an audit trail.
The retrieval core above is what makes that work, and the most ambitious rebuilds run on Iris as part of an Orchid engagement.
Iris is built, deployed, and run by Nomad on per-customer isolated infrastructure. It usually starts with an Operational Audit so we model the real process before we build it.
Three sizes, no guesswork.
Every engagement starts with a Discovery, and the Discovery fee is credited against the Build if you go ahead. We don’t quote on faith.
Stakeholder interviews, process mapping, technical feasibility, model selection. You leave with a written recommendation sized small, medium and large, and a straight answer on whether to proceed.
Fixed scope, fixed price, weekly Friday demos. Production-grade auth, audit logging, monitoring and a model-agnostic AI layer from day one. 30 days of post-launch support included.
For a need that is structural rather than a single workflow: multiple teams, real integration with your stack, phased rollout. 90 days of managed post-launch support.
Build your operating system.
A 30-minute call about how your business runs today, and what Iris would look like shaped around it.
Talk to us →