Contact us

How we work

Diagnose. Design. Implement. Sustain.

One method for procurement, supply chain and the AI that now runs inside both: facts first, an operating model before any platform, delivery on site, and routines that outlast the engagement.

How we work

Facts first. Operating model before platform. Delivery on site.

One method across procurement, supply chain and AI, from the diagnostic to the routines that outlast the engagement.

  1. Diagnose

    Spend, flows, inventory and process data from the systems. A value case leadership can sign.

  2. Design

    Operating model, category and network strategies, platform blueprint and the AI portfolio.

  3. Implement

    Sourcing waves, go-lives, planning cycles and the first agents in production, on site.

  4. Sustain

    Routines, KPI reviews and trained teams that run without us.

  1. Diagnose

    3–6 weeks

    What is really happening, and what is it worth?

    Spend, flows, inventory, service and process data extracted from the systems, not from slides. Interviews across sites and functions. A maturity assessment against a practical reference model, and a value case leadership can sign.

    What this involved

    • Spend cube from purchase orders and invoices, one taxonomy across sites
    • Process mining on Procure-to-Pay and order-to-delivery
    • Inventory, service and cost-to-serve baseline
    • Supplier and network risk mapping
    • Operating-model and capability assessment
    • Data readiness for automation and AI

    Impact

    • Fact base and baseline KPIs
    • Value case by lever, stated as ranges with assumptions
    • Prioritised roadmap by wave
    • Quick wins launched during the diagnostic

    AIWhere AI can help is decided here, from the data and the decisions that actually exist, not from a vendor demo.

  2. Design

    6–10 weeks

    What should the operating model be, and what will run it?

    Target operating model, category strategies, network and planning design, platform blueprint and the AI use-case portfolio, designed with the people who will own them and validated in workshops with executive leadership.

    What this involved

    • Procurement operating model: mandate, category governance, decision rights, KPI language
    • Category strategies and sourcing waves
    • Network and inventory design with scenario modelling
    • S&OP / IBP cycle, roles and calendar
    • Source-to-Pay and planning platform blueprint
    • AI use-case portfolio with decision rights and controls per use case

    Impact

    • Target operating model and organisation
    • Category and network strategies
    • Platform configuration blueprint
    • AI portfolio, data foundation plan and governance charter

    AIEach AI use case is designed as a decision with an owner, a threshold and a fallback, inside the KPI language agreed with Finance.

  3. Implement

    by wave

    Does it work on the shop floor, in the plant, in the system?

    Sourcing waves, platform activation, network moves, planning cycles and the first AI agents in production, delivered on site with the client’s teams. OREDJA leads as consultant, interim manager or programme director, whichever the situation requires.

    What this involved

    • Sourcing waves and contract activation
    • Source-to-Pay and planning platform configuration and go-live
    • Network, warehouse and line-feeding changes
    • S&OP cycle running with real decisions
    • AI agents in production with humans in the loop
    • Training, routines and shop-floor management

    Impact

    • Ratified savings and service improvements
    • Live platform and clean master data
    • Operating routines and trained teams
    • Agents in production with audit trails

    AIAgents go live one workflow at a time, with exception handling designed before automation and every decision traceable.

  4. Sustain

    ongoing

    Will it still work when we are gone?

    Performance routines, KPI reviews, capability development and a hand-over that makes the organisation independent. OREDJA stays available for reviews, next waves and the governance of the AI that keeps running.

    What this involved

    • Daily and monthly performance routines
    • KPI reviews with Finance and operations
    • Capability building and succession
    • Continuous improvement of category and planning practices
    • AI model and agent monitoring, drift and control reviews
    • Next-wave planning

    Impact

    • Documented routines and owners
    • KPI dashboards the business actually uses
    • Governance calendar for AI systems
    • Independent teams

    AIAI systems are reviewed like any other control: monitored, audited and retired when they stop earning their place.

How we work

Turning complexity into opportunity.

One method for procurement, supply chain and the AI that now runs inside both: facts first, an operating model before any platform, delivery on site, and routines that outlast the engagement.

ProcurementSupply ChainAI OrchestrationImpact01 → 0401Diagnose3–6 weeks02Design6–10 weeks03Implementby wave04Sustainongoing

What is really happening, and what is it worth?

Spend, flows, inventory, service and process data extracted from the systems, not from slides. Interviews across sites and functions. A maturity assessment against a practical reference model, and a value case leadership can sign.

Impact

  • Fact base and baseline KPIs
  • Value case by lever, stated as ranges with assumptions
  • Prioritised roadmap by wave
  • Quick wins launched during the diagnostic

AIWhere AI can help is decided here, from the data and the decisions that actually exist, not from a vendor demo.

Select an element to see what it means in practice

AI Orchestration

How we implement AI in procurement and supply chain

The method is the same; the discipline is stricter. AI is introduced where an operating model, clean data and clear decision rights already exist, and it is measured in the numbers the business already uses.

  1. 01

    Select by decision, not by technology

    Use cases are ranked by the value and frequency of the decision they improve: supplier selection, order exception, forecast override, contract clause, risk alert. Chatbots do not make the list on their own.

  2. 02

    Data foundation before agents

    Classified spend, clean item and supplier masters, reliable inventory records and a documented process are the foundation. Without them an agent automates guesswork.

  3. 03

    Decision rights and humans in the loop

    Every automated step has an owner, a threshold above which a person decides, and an exception path. Agents propose and execute within limits; people stay accountable.

  4. 04

    Governance aligned with ISO/IEC 42001

    An AI management system that covers purpose, risk, data, monitoring and audit trails, sized for the organisation rather than copied from a standard.

  5. 05

    Measured in the KPI language of the business

    Cycle time, first-time-match rate, forecast accuracy, inventory, cost avoidance, savings ratified by Finance. If the number does not move, the use case is stopped.

  6. 06

    One workflow first, then waves

    The first agent goes live on a single, measurable workflow. Each following wave reuses the same data foundation, controls and governance, which is what makes scale safe.

The first ninety days of an AI orchestration programme

A realistic sequence, drawn from how the flagship programme runs today.

Weeks 1–2

Baseline and data readiness

Process data, master data quality and current decision points documented. The first backlog of exceptions is measured, not estimated.

The first ninety days of an AI orchestration programme

Platforms and tools we work with

OREDJA is independent of vendors. These are the environments in which our teams have designed, configured or run procurement and supply chain operations.

Procurement

  • Coupa
  • SAP Ariba
  • Ivalua
  • Jaggaer
  • SAP S/4HANA

Supply chain design and planning

  • Coupa Supply Chain Design
  • SAP IBP
  • Kinaxis
  • o9 Solutions

AI, data and orchestration

  • OpenAI
  • Claude
  • Mistral AI
  • Palantir
  • Databricks
  • Microsoft Azure AI

Governance

  • ISO/IEC 42001
  • EU AI Act
  • IAPro.ai

Questions we are asked before an engagement

How long before we see results?

Quick wins are launched during the diagnostic, typically within the first six weeks. Ratified savings from a sourcing wave take three to six months; a network or planning redesign shows in inventory and service within two to four quarters.

Do you implement, or only advise?

Both, and the second is worth little without the first. OREDJA’s leadership has run procurement and supply chain functions as interim managers and executives. We stay through go-live and the first months of operation.

We already have a platform. Why an operating model?

Because the platform records decisions; it does not make them. Most dormant sourcing and contract modules are dormant for lack of an owner, a KPI and clean data, not for lack of software.

Is AI realistic for a mid-sized organisation?

Yes, if it starts with one measurable workflow on clean data. The organisations that struggle are those that start with a platform-wide promise. We start with an exception queue.

How do you work with our teams?

On site, embedded, with your people owning the routines from the first week. Handing over a running organisation is the point of the engagement.

Let’s discuss

What could stronger procurement and supply chain do for your business?

Discuss your priorities

No commitment. Just a conversation.

  • Global experience
  • Local understanding
  • Lasting impact