Collaboration Portal · Forvis Mazars × TPO Management Services

01 Executive Overview

Context

From workshop discussion to a living operational blueprint

TPO Management Services operates in a complex operational environment: Oracle-backed systems, air-gapped infrastructure, GPU clusters, and a growing need to consolidate operational intelligence across facilities, logistics, finance, and decision support.

The opportunity is to make TPO Management Services AI by design: connecting every system, every process, every decision into one Autonomous Intelligence Network.

The next phase moves from workshop discussion to a structured collaboration: co-designing AIN, an open-source operational intelligence platform built around TPO's actual technology strategy.

Oracle integration Air-gapped infrastructure GPU clusters Operational intelligence Open-source first No vendor lock-in

About this portal

Participants

Aymen Al Saadi, Chief Digital Officer
Ali Al Khumairi, Director of Procurement
Forvis Mazars advisory team
NaasAI / ABI technical team

Confidentiality

Strictly confidential for named recipients.
Hosted on a private server with password access.
Do not share credentials or forward links externally.

02 AIN Platform

Platform concept

AIN: the operational intelligence platform for TPO

AIN is the operational platform. ABI is the open-source operating system backbone that powers its agents, workflows, ontologies, integrations, and knowledge graph.

The same logic as BOB for Forvis Mazars: a platform built on open-source foundations, adapted to the operational context of the organisation, owned and controlled by the organisation.

Executive workspace Operational command layer Knowledge graph AI agents Workflow orchestration Decision support Traceability Auditability

Platform capabilities

Executive workspace: unified view across operations, briefings, and decisions.
Operational command layer: processes encoded, tracked, and improvable.
Knowledge graph: semantic spine linking vendors, assets, staff, contracts.
AI agents: human and machine agents working alongside each other.
Workflow orchestration: procurement, approvals, reporting, incident response.
Decision support: intelligence briefings, risk alerts, spend patterns.

Operational Use Cases

03 Operational Intelligence

Practical intelligence across the S1-S9 operational framework

Starting with the highest-value operational domains for TPO. Each use case is a concrete, deployable workflow built on the AIN platform and powered by the ABI operating system backbone.

S3 · Operations

Facilities management and work orders
Service requests and task coordination
Event planning and scheduling
Incident tracking and escalation

S4 · Logistics

Vendor management and procurement workflow
Inventory tracking and maintenance scheduling
Contractor performance monitoring
Asset lifecycle management

S8 · Finance & Contracts

Purchase approvals and straight-through processing
Contract visibility and lifecycle tracking
Invoice matching and budget follow-up
Spend pattern analysis

S2 · Intelligence

Executive briefings and situational awareness
Risk alerts and reputation monitoring
Market and supplier intelligence
Cross-domain signal aggregation

View procurement decision flowchart

04 Open Source Architecture

Technology

Open-source first, model-agnostic, infrastructure-sovereign

ABI is the open-source operating system backbone powering AIN. Every component is replaceable. No dependency on a single model provider, no SaaS lock-in, no data leaving the organisation without explicit control.

AIN is the platform experience. The stack underneath is fully open and auditable.

Open-source first Model-agnostic Air-gapped compatible Infrastructure sovereign No vendor lock-in Deployable on-premises GPU cluster ready

Materials

Open source AI stack · slides Knowledge graph

05 Air-Gapped Deployment

Infrastructure

Runs inside TPO infrastructure. No external dependency.

AIN is designed to run entirely within TPO-owned infrastructure. No SaaS dependency. No data leaving the perimeter without explicit consent. Local models, controlled data flows, integration with Oracle and internal systems.

Deployable on TPO's existing servers and GPU clusters. Designed for security, privacy, and full operational control.

On-premises deployment Oracle integration Local model inference Controlled data flows GPU cluster support Zero external SaaS dependency

Materials

Sovereign AI architecture · slides Deployment architecture brief

06 Initial Workshop Materials

Reference materials

Vivatech Paris 2026 · Initial Workshop Deck

These materials were prepared as a first discussion draft for the Vivatech Paris 2026 session and are being updated based on new information from TPO. They represent the starting point of the collaboration, not the current agreed direction.

The consolidated workshop deck covers the foundational topics introduced at Vivatech Paris: sovereign AI architecture, open-source stack, procurement workflows, and the AI by Design Flywheel. Use as reference alongside the evolving AIN collaboration.

Sovereign AI Open-source stack Procurement workflow AI by Design Flywheel S1-S9 framework Data platforms

Materials

Full workshop deck Procurement use case Knowledge graphs

07 Roadmap

Next steps

From initial workshop to operational platform

A four-step path from alignment to a live operational intelligence platform. Each step builds on the previous and produces a concrete deliverable.

1

Alignment Workshop

Validate priorities, systems, constraints, and deployment environment. Define the scope of AIN Phase 1.

2

Operational Blueprint

Map processes, systems, data sources, and first AIN use cases. Produce a structured design for S3 and S4 workflows.

3

AIN Alpha

Build the first operational graph and workspace around S3 facilities and S4 procurement. Live inside TPO infrastructure.

4

Enterprise Rollout

Expand toward full S1-S9 operational intelligence. Oracle integration, executive briefing layer, and continuous improvement loop.