Home / Case Studies

Selected work

AI that solved real problems, and what changed.

Selected engagements across retail, consumer goods, technology, manufacturing and supply chain. Client identities stay confidential unless a client has agreed to be named.

FeaturedRetail · AI optimisation

Intelligent Retail Optimization Platform

An optimisation layer on top of traditional planogram tools. It decides which products belong in each store and how much shelf space each one deserves, then generates the planogram automatically.

  • Store clustering by demand pattern and customer profile
  • Assortment optimisation with Keep, Delist and Consult decisions
  • Space optimisation for maximum revenue per square foot
Read the full case study
+8.3%Revenue per store
71 → 89%Shelf utilisation
$2.5MEstimated annual revenue gain
16 h → 30 sPlanogram creation time

Pilot results for an urban, high-income store cluster.

AI for Commercial Decisions

Measurement, forecasting, GenAI and allocation
Global consumer goods · Retail

Enterprise-Scale Causal Measurement Framework

Commercial investments across markets had no unified incrementality measurement. Leaders relied on correlation-based reporting, limiting clarity on true ROI.

See solution & impactHide details
Solution
  • Scalable causal inference architecture combining experimental design, synthetic control and quasi-experimental models
  • A repeatable measurement system adaptable across markets
  • Governance guardrails and standardised reporting
Impact
  • Data-driven reallocation of commercial budgets
  • Higher decision confidence at regional and global level
  • Incrementality established as a decision principle
Retail & apparel

Demand Forecasting & Bias Remediation

Volatile forecast error across product hierarchies led to excess inventory, stockouts and poor allocation.

See solution & impactHide details
Solution
  • Hierarchical and probabilistic forecasting pipelines
  • Promotional and event-level drivers built into the models
  • Forecast error decomposition to find systemic bias
  • Monitoring dashboards for sustained accuracy
Impact
  • More stable forecasts across categories
  • Lower excess inventory exposure
  • Greater cross-functional trust in planning outputs
Consumer products · Multi-channel

Constrained Optimisation for Resource Allocation

Inventory, budget and capacity decisions were made with rules of thumb instead of formal optimisation.

See solution & impactHide details
Solution
  • Constrained optimisation models covering service levels, cost structures and demand uncertainty
  • Scenario simulation environment for leadership
  • Decision-support tools embedded in recurring planning cycles
Impact
  • Better allocation efficiency within budget
  • Less manual decision variability
  • Transparent trade-offs for senior leaders
Technology · Knowledge operations

Enterprise GenAI Workflow Integration

Manual document analysis and insight synthesis slowed operational workflows.

See solution & impactHide details
Solution
  • Governed GenAI workflows with retrieval, structured reasoning and evaluation layers
  • Human-in-the-loop validation
  • Cost and performance monitoring
Impact
  • Shorter cycle times for analytical work
  • More consistent outputs
  • A scalable base for future automation
Multi-market enterprise

AI Operating Model & Capability Architecture

AI initiatives were fragmented, with no shared governance or long-term capability roadmap.

See solution & impactHide details
Solution
  • AI operating model across central and market teams
  • Roles, governance principles and delivery standards
  • A phased AI roadmap tied to business value streams
Impact
  • Clearer execution across regions
  • Faster move from pilot to production
  • Stronger executive AI decision-making

AI for Operations & Manufacturing

Production, maintenance and quality
Manufacturing · Industrial operations

Production Optimisation & Capacity Intelligence

Production planning was constrained by capacity limits, demand swings and cost pressure. Manual planning led to inefficiency and constant reactive changes.

See solution & impactHide details
Solution
  • Production optimisation models covering machine capacity, shift constraints and material availability
  • Demand forecasts linked directly to plant-level scheduling
  • Scenario simulation tools for capacity expansion planning
Impact
  • Production better aligned with demand signals
  • Less idle capacity and fewer bottlenecks
  • Structured, long-term capacity planning decisions
Manufacturing · Asset-intensive plants

Predictive Maintenance & Asset Reliability

Unplanned downtime caused disruption and cost overruns. Maintenance was reactive or calendar-based instead of condition-driven.

See solution & impactHide details
Solution
  • Predictive models built on equipment telemetry and historical maintenance logs
  • Failure-probability estimation framework
  • Risk-based maintenance prioritisation dashboards
Impact
  • Lower exposure to unexpected downtime
  • More efficient maintenance scheduling
  • Clear reliability visibility for plant leadership
Industrial manufacturing

Quality Intelligence & Anomaly Detection

Quality deviations were caught late in the production cycle, driving up rework and scrap.

See solution & impactHide details
Solution
  • Anomaly detection across multivariate production signals
  • Real-time monitoring framework
  • Root-cause analysis workflows
Impact
  • Process deviations identified earlier
  • Less defect propagation down the line
  • Better coordinated response across functions
Manufacturing · Elevators & escalators

Smart Manufacturing Control Tower

Plant teams needed one live view of orders, lines and dispatch, and early warning of delays and shortages, on top of existing SAP systems.

See solution & impactHide details
Solution
  • Real-time control tower dashboards and shop-floor operator screens
  • Six AI engines: delay prediction, plan optimisation, material shortage, line utilisation, OEE and dispatch
  • All engines reading from existing SAP through OData and RFC APIs
Impact
  • One shared, live picture of plant performance
  • Delays and shortages visible days in advance
  • No change to the client's core SAP landscape

Supply Chain & Network Optimisation

Inventory and distribution design
Global supply chain · Distribution

Multi-Echelon Inventory Optimisation

Inventory buffers across distribution tiers were misaligned: excess holding cost in some nodes, stockouts in others.

See solution & impactHide details
Solution
  • Multi-echelon inventory model under demand uncertainty
  • Service-level and cost trade-offs optimised across nodes
  • Scenario tools for network evaluation
Impact
  • Better service levels with less working capital
  • Less network-wide variability in stock positions
  • Cost-service trade-offs quantified for leadership
Manufacturing · Retail distribution

Distribution & Network Design Simulation

Network expansion and redesign decisions relied on static assumptions with no quantitative scenario testing.

See solution & impactHide details
Solution
  • Simulation models for distribution flow and capacity stress testing
  • Alternative node placement and routing strategies evaluated
  • Cost, service and capacity constraints combined
Impact
  • Better long-term infrastructure decisions
  • Fewer structural inefficiencies in distribution
  • Executive-ready decision scenarios
Your plant next

Have a problem like these?

Share a short description and a sample of your data. We'll come back with how we would approach it.