Skip to content
NOVO/MODUS
HOW WE WORK

The
Modus

Modus is Latin for method. Ours has five stages that run as a loop, not a line — because a business never stops changing, and neither should the systems that run it.

01

Discover

Understand the organisation, its systems, workflows and constraints.

We spend time where the work happens. Not reading requirements — watching orders move, sitting with the people who handle exceptions, tracing data between systems. The output is a map of how things actually work, which is almost never how the org chart says they work.

WE ASKWhere does time go? Where does information get re-typed? What breaks, and who fixes it by hand?
YOU GET
  • Current-state map of systems and workflows
  • Ranked list of friction and risk
  • A shared vocabulary for the problem
02

Design

Simplify the problem and define the right architecture.

Before any tool or framework is chosen, we design the target way of working: fewer steps, clear ownership of data, and the smallest architecture that supports it. Most of the value of a project is decided here, on paper, where changes are cheap.

WE ASKWhat can be removed entirely? What must stay human? What is the single source of truth for each kind of data?
YOU GET
  • Target-state process design
  • Architecture and integration decisions, written down
  • Sequenced roadmap with a first thin slice
03

Build

Engineer reliable software and integrations.

We ship a working end-to-end slice within weeks and grow it with real users. Tests, observability, infrastructure as code and documentation are part of "done", not a later phase. The code is readable by the next team, because there is always a next team.

WE ASKIs this the simplest thing that works? Can we explain every decision? Could someone else run it tomorrow?
YOU GET
  • Production software, API-first
  • Automated tests and CI/CD
  • Documentation and decision records
04

Automate

Remove repetitive work and connect systems.

With the core in place, we take the repetitive majority of the work away from people and give it to the system — hand-offs, synchronisation, reporting, routine decisions. Exceptions go to a person, with full context, so judgement is spent where it matters.

WE ASKWhich steps exist only because two systems don't talk? Which decisions are actually rules? What should never be automated?
YOU GET
  • Integration layer with retries, logging and alerts
  • Exception routing and runbooks
  • Measured reduction in manual work
05

Evolve

Measure, optimise and continuously improve.

We measure against the baseline from Discover and keep adjusting. Systems drift, businesses change, and the cheapest moment to simplify is early. Evolve is where AI usually enters — once there is clean data and a clear decision to improve.

WE ASKDid it work? What changed since we started? What has become the new bottleneck?
YOU GET
  • Outcome report against the original baseline
  • Optimisation backlog, prioritised by impact
  • The questions for the next Discover
WHY A LOOP

Evolve feeds Discover.

A linear process assumes the problem holds still. It does not. The measurements from Evolve become the questions for the next Discover — sometimes a small adjustment, sometimes a new project, occasionally the decision that something should be switched off.

This is also why we prefer continuous partnerships to hand-over-and-disappear. The people who built the system are the best people to keep it simple.

Start with Discover
COOKIES

We use only strictly necessary cookies. Optional analytics load only if you agree. Preferences