Start here

Advisory first, and most often your own twin.

Novus-i2 is the firm. NovusLens is the product the firm builds and offers. An engagement usually starts as advisory work and often ends with the customer running its own NovusLens twin. Advisory work continues alongside the twin, helping people interpret what it shows.

Advisory work usually comes first, the customer often comes to run its own twin, and advisory judgement continues alongside it.

Why the engagement starts with advisory

Advisory work supplies the judgement that a tool does not supply on its own. An adviser examines how people make decisions and who gains or loses from a change. They help explain what differences between written accounts and recorded activity mean for the people involved. The product shows the evidence. Interpreting it in the customer’s setting is the advisory part.

Today the main deliverable is advisory work. The main output of an engagement at this stage is the Agentic Readiness Assessment. It can be accompanied by an initial Accountability Map, which sets out the Capability, Authority and Responsibility of the customer’s AI agents.

How the customer comes to run its own twin

Most engagements are designed to end with the customer running its own twin, NovusLens’s up-to-date model of how the organisation is wired. The twin is built from the customer’s own records. The customer keeps the records and the twin, and the team that uses it is the customer’s team.

A twin runs on a Docker host the customer controls. The operator runs ./lens setup, which asks a few questions, creates the secrets and starts the installation with Docker Compose. The database runs inside the installation, and so can the model server. Setup checks each setting before it starts anything. When a check fails, setup names the setting and the fix, and starts nothing.

Installation gives the steps. The operating section covers running the installation after that.

Advisory continues alongside the product

Once the customer runs its own twin, Novus-i2 does not withdraw. The advisory practice stays alongside the product to help interpret the evidence. It reviews what the twin shows, helps frame a management question, and supports the people who must decide. The twin cannot make those judgements, so the two work together.

NovusLens is not a consultancy, and it is not a system of record for the customer’s assets. It maps and advises. It does not run the customer’s systems or agents.

Where the product stands

NovusLens is being built in stages. The first stage has passed our own acceptance checks. The second stage is built and will be checked in-house next. Later stages continue to be built. Engagements today are advisory-led, and the product moves towards the customer running the twin as those stages are completed. This site describes the product as it is built. It does not give a date for when each part will be ready for a customer.

What a demonstration looks like

A demonstration runs on an installation made with ./lens setup, and takes about twelve minutes. It asks one question on behalf of a fictional sponsor, and it uses a fictional organisation called Meridian Grid. Every person and company in that organisation is invented, and every source in the demonstration pack declares itself simulated. Some steps show agent activity that is generated demonstration evidence, not records from a real runtime. The presenter says so before running it.

In a real engagement, each record comes from the customer’s own systems, through a source the customer has enrolled. The demonstration is a way to show the product. It is not a claim about the customer’s organisation.

What to read next

If you are new to the product, take your first hour. If you want the reasoning behind the measures, read the concepts section first.