Fathy GadDesign Automation · Enterprise AI
Executive briefing

Case studies.

Three programmes, each set out the way a board would review them: the business problem, the architecture chosen and what changed for the organisation.

Case 01 · Middle East & India

Building an enterprise automation capability

Multidisciplinary: structure, architecture, MEP, ports & marine

Challenge

Fee pressure and fewer tenders, while skilled engineers spent a material share of their hours on modelling, scheduling and documentation.

  • Isolated scripts with no shared deployment, measurement or ownership
  • Know-how held by individuals rather than the organisation
  • No evidence base for technology investment decisions

Strategic architecture

A layered platform, deployed and governed like a product — not a collection of tools.

  • Add-ins inside the design applications teams already use
  • A sync layer that ships updates and collects usage telemetry
  • Central application and data services, with an AI layer on the same data
System architecture
  1. Design add-insInside Revit, Rhino and CSI tools
  2. Sync & telemetryUpdates out, usage in
  3. App & data servicesAPIs, logs, licensing
  4. AI layerAgents on shared data
  5. Value ledgerHours returned by project and discipline
Enterprise impact
Production systems used daily by multidisciplinary teams
Updates shipped centrally; usage measured on every run
Know-how captured in reusable tools, not individuals
Technology investment backed by usage evidence
Executive takeaway

Automation becomes a capability — not a cost — when it is deployed, measured and owned like a product.

Case 02 · The Mukaab · Al Maktoum International Airport · NEOM Vault

Computational delivery at mega-project scale

Saudi Arabia and the United Arab Emirates

Challenge

Landmark geometries where every design change ripples through hundreds of thousands of elements across BIM and analysis models.

  • Manual re-modelling could not keep pace with design iteration
  • Translation errors between modelling and analysis
  • Programme risk from late coordination

Strategic architecture

Scripted model generation, with BIM and structural analysis kept in sync from a single source of design rules.

  • Parametric engines generate structure directly from design rules
  • Live integration between BIM and analysis environments
  • A design change propagates in one run instead of a re-model
System architecture
  1. Design rulesGeometry, grids, constraints
  2. Parametric engineScripted generation
  3. BIM modelCoordinated and documented
  4. Analysis modelBuilt and loaded automatically
  5. Drawings & quantitiesRegenerated on change
Enterprise impact
Design changes propagated in one run, not a re-model
BIM and analysis models kept in sync
Fewer translation errors between environments
Coordination issues surfaced earlier in the programme
Executive takeaway

At this scale the advantage is the speed at which a design can change — not the speed at which it can be drawn.

Case 03 · Middle East

A governed enterprise AI layer for engineering

Design, BIM, structural analysis and resource planning

Challenge

Generic AI tools hallucinate on engineering data, and unmanaged adoption creates quality and liability exposure.

  • Project knowledge scattered across documents, models and email
  • Slow responses to RFIs, comments and code queries
  • No control over how AI outputs enter deliverables

Strategic architecture

Data first, then agents, then a human gate — AI that drives proven tools rather than improvising.

  • Uniform data schema and centralised, indexed design data
  • One shared data and agent layer for BIM and structural design agents
  • MCP connectors to existing automation, with mandatory engineer approval and a full audit trail
System architecture
  1. Uniform data schemaIndexed by project and model
  2. Shared data & agent layerRAG, MCP connectors
  3. AI agentsBIM agent · Design agent
  4. Engineer approvalMandatory, never skipped
  5. DeliverableTraceable to source
Enterprise impact
Every output reviewed and approved by an engineer
Answers traceable to source documents
Proven automation reused instead of improvised
Full audit trail of runs and changes
Executive takeaway

AI earns its place in engineering delivery through governance, not demonstrations.

Discuss this with Fathy Back to home