In active use

Custom analytical tooling

Builds the model, calculator or interface an institution needs and does not have — and, critically, builds it so the institution can still run it after the assignment ends.

What it looks like

Custom build — chosen for who operates it

WorkbookFor a team that lives in Excel

ApplicationWhere there is engineering capacity

InterfaceBuilt with Rubik

HandoverDocumented method · version control · training · the client owns it

Schematic layout — the tool's real structure and outputs, drawn rather than screenshotted so no client data appears. Bars and blocks are placeholders, never figures.

Decision supported

Whatever recurring decision the institution has to make repeatedly and currently cannot make on its own evidence.

Intended user

Institutions that keep re-commissioning the same analysis because the last consultant left with the spreadsheet.

Inputs

The decision that has to be made, and how often
Data the institution already holds, in whatever state it is in
Who will operate the tool and what they can realistically maintain
Reporting or regulatory obligations the output must satisfy

Outputs

A working tool — spreadsheet model, Python application or web interface, chosen for who has to run it
Documented method, so results can be defended and reproduced
Handover and training for the team that will own it
Version control and a change log where the tool will be maintained over time

Methodology summary

We build in the format the client can actually sustain. A ministry team that lives in Excel gets a rigorous, documented workbook rather than a Python service nobody can maintain; a team with engineering capacity gets an application with version control and tests. Where the interface needs real software engineering, our partner Rubik builds it. The choice is made on who operates it afterwards, never on what is most impressive to deliver.

Limitations

We do not sell software licences or run a SaaS product. These are commissioned tools, built for one institution's decision and owned by it.
A tool cannot fix a data problem. Where the underlying data does not exist, the first deliverable is an honest account of that gap, not an interface that hides it.
Sustained maintenance is the client's, unless a support arrangement is agreed separately. We build for handover rather than for dependency.

Screenshots

Custom analytical tooling — interface rendering showing the executive dashboard with a demo dataset. — Interface rendering — the tool's real screens and outputs, produced as a design rather than captured from a running session. Demo dataset throughout: no client name, real place or real figure appears. **Deliberately cropped to the results pane.** The navigation column is not shown: what a tool produces is the useful thing to publish, while its full module structure is method rather than proof. Cropped, not blurred — a blurred screenshot reads as either a mock-up or as something being hidden badly, and both cost more than the architecture is worth.

Access model

Commissioned. Scoped and priced as part of an engagement or as a standalone build.

Request a demonstration

Tell us the context, timeline and decision you're working toward.

Request a demonstration

We reply within two business days.