Business areas

Banking and essential services

A flag that cannot be explained cannot be contested, and therefore cannot be used.

The environment

The quality of the input data decides the value of the system, before any analysis.

An analytical flag must be capable of being explained, contested and corrected. The quality of the input data governs the value of the system more than the sophistication of the analysis: a model fed inconsistent data produces results nobody can defend to the person they point at.

Sensitive data imposes a second discipline. Health data calls for the strictest minimisation and partitioning by purpose; continuity of care takes precedence over any consideration of administrative convenience. A system that simplifies administrative work by widening access to records has solved the wrong problem.

The third constraint is distribution. A branch network accumulates installations built at different times, none documented in the same way. A link failure then isolates a branch, and head office learns of it only through a customer call; the real cost of heterogeneity shows in travel expenses, not in the equipment invoice.

Dominant constraints

What this environment imposes

A dominant constraint governs architecture, partners, equipment and method more than any list of services does.

Finance and fraud prevention

Dominant constraintData quality and contestability of results

An analytical alert must be explainable, contestable and correctable. The quality of the input data governs the value of the system more than the sophistication of the analysis does.

Health and essential services

Dominant constraintData minimisation and continuity of care

Health data calls for the strictest minimisation and for separation by purpose. Continuity of care takes precedence over any consideration of administrative convenience.

Evidence

What we have delivered in this environment

All projects

Before committing

Questions to settle internally

An organisation that cannot answer these questions is not ready to launch the programme. Saying so is better than selling it a study.

  • Can a result produced by the system be explained to the person it points at?
  • What data is collected, for what declared purpose, and for how long?
  • Which branches cannot tolerate an interruption, and for how long?
  • Does head office learn of a failure before the first customer call?
  • Who, in house, must be able to create and withdraw an access without going through the provider?
  • What downtime does the executive accept, and has it been written down?

Consultations · Pre-qualifications · Partnerships

Let us discuss the actual case.

Describe the site, the dominant constraint and the deadline. We will say what requires a preliminary study and what can be committed directly.