Skip to content

Capabilities

Software engineering is the delivery medium. The work is computational: formulating operational problems, solving them, and putting the result into systems that run.

  1. 01 / Optimise

    Find better decisions inside real constraints.

    Scheduling, allocation, routing, capacity, inventory, matching and sequencing are combinatorial problems wearing operational clothes. We formulate the constraints the business actually has — labour, geography, contracts, assets, service levels — and engineer solvers and decision engines that produce feasible, high-quality plans.

    • Workforce, fleet and asset allocation
    • Scheduling and sequencing under rules
    • Routing, logistics and network design
    • Inventory, procurement and capacity planning
    • Multi-objective trade-offs, not a single convenient metric
  2. 02 / Model

    Understand the system before changing it.

    Expensive operational change is often committed on the basis of intuition. Modelling and simulation make the behaviour of a process visible: bottlenecks, queueing, utilisation, failure modes, demand variability and the consequences of a proposed policy.

    • Discrete-event and Monte Carlo simulation
    • Capacity, demand and scenario models
    • What-if and stress analysis
    • Operational digital twins where they earn their keep
    • Sensitivity of decisions to uncertain inputs
  3. 03 / Engineer

    Put domain intelligence into production software.

    The differentiator is not another interface over a database. It is software that contains the rules, models and algorithms of the operation — platforms, decision engines, control surfaces, integration layers and APIs that other systems can call.

    • Internal operational platforms
    • Decision and rules engines
    • Workflows with computational steps, not only human routing
    • Data pipelines that feed decisions rather than dashboards
    • Integration with existing systems of record
  4. 04 / Improve

    Measure what the system actually did, then iterate.

    A decision system that is never confronted with outcomes will drift. We instrument the live operation, compare intended decisions with what occurred, and refine models, constraints and software as the organisation learns.

    • Feedback from live operational data
    • Exception handling and human-in-the-loop control
    • Progressive tightening of models and rules
    • Operational metrics tied to decisions, not vanity charts

Optimisation, in operational language.

A typical problem is not abstract. It looks like hundreds of jobs, tens or hundreds of resources, geography, labour rules, contractual windows, service levels and a cost or utilisation objective. The number of conceivable assignments is enormous. People currently resolve it with experience, overtime and exception handling.

Industry Analytics translates those constraints into a structure a computer can search — then builds the software that produces plans operators can actually use. Linear and mixed-integer programming, constraint programming, graph algorithms, heuristics and custom methods are tools. The client-facing result is a feasible decision, on time, in the working system.

Methods are chosen for the problem. Deterministic optimisation, simulation, statistical models and machine learning are used where they are warranted — not as a default costume.

Constraint field

Automation, held in proportion.

Where a decision can be executed with a human in the loop, or where documents and exceptions need triage, we will automate. Agentic and machine-learning techniques are one class of computational tool among others. They do not define the firm.

Representative problem types

Framed as problem classes, not completed engagements.

  • Scheduling

    Time, resources and rules that cannot all be satisfied by a spreadsheet.

  • Allocation

    Assigning scarce people, assets or budget under competing demands.

  • Routing

    Movement of work, goods or people through a constrained network.

  • Capacity

    What the system can absorb, where it saturates, and what to change first.

  • Planning

    Medium-horizon decisions that couple inventory, labour, production and service.

  • Forecasting

    Demand, workload and risk as inputs to a decision — not as decoration on a chart.

  • Workflow

    Processes whose structure, not only their interface, needs to be redesigned.

  • Operational control

    What should happen next when the live system diverges from the plan.

  • Simulation

    Testing a change in a model before spending it in the real operation.

  • Bespoke systems

    Software the organisation must own because no product encodes its logic.

If the problem is operationally hard, start there.

A useful conversation begins with the decision that is currently made by heuristic, spreadsheet or committee — and with the constraints that make it difficult.