Super charging Computational Design (CoDe)

An engineering view on AI and coding efficiency

1. A short history of CoDe in engineering

Computational Design started as a way for engineers to describe logic to automatically generate instead of drawing repeatable or novel geometry. It allowed relationships between loads, spans, and constraints to be defined mathematically so that computers could handle repetition and updates automatically.

Tools like Grasshopper and Dynamo have made this possible. They connect design variables to algorithms that can modify an entire model through one change. They changed how engineering staff approached design, particularly in complex, multi-disciplinary projects.

But the process requires deep technical skill and mastery. Each platform has its own rules, language, and data structure. Currently, only a small group of specialists use them effectively.

CoDe is powerful but remains a niche capability. Until now!  The gap is now being closed by AI-driven automation and well trained coding assistants.  I call this ‘AI Led Design’.

2. Coding at engineering speed

Artificial Intelligence has made coding faster and easier to use. Engineering Coding Platforms (ECPs), together with AI assistants, allow engineering staff to produce working code quickly without spending hours on setup or debugging.  Obviously, know-how and quality are key!

An engineer can now describe a task for example, “create a Python function to generate a beam schedule for various spans and reinforcement grades” and receive structured, working code almost instantly. Syntax and structure are handled automatically, leaving the engineer to review the logic and results.

ECPs extend this by allowing engineers to package scripts into usable design tools with input forms, 3D outputs, and version control. This turns code into an engineering asset that can be reused across teams instead of living on one person’s computer.

These systems don’t replace engineering expertise at all; they speed up the parts of coding that used to slow everyone down.  It might even disrupt conventional tools currently use in CoDe.

The major transformation is that CoDe moves from automation using scripts to full on design, with engineers being assisted by intelligence!  #GameChanger!

3. Making optioneering intelligent again

In most projects, optioneering stops once deadlines tighten. After one feasible solution is found, exploration ends. With AI supported Computational Design, testing multiple options becomes practical again.

Once rules such as load limits, spacing, or material properties are defined, ECPs and AI coding tools can automatically generate and evaluate design variants. Engineers can compare options based on performance, cost, or sustainability criteria within minutes rather than days.

This brings back genuine engineering exploration guided by rules, validated by logic, and delivered at the speed of computation, at any time.  The present situation where everyone needs to freeze ‘information requirements’ VERY early on in a project phase becomes a thing of the past.

4. From specialist task to common practice

For years, CoDe was limited to a few experts who wrote scripts for everyone else. That created bottlenecks and slowed adoption, all unintentional. AI based coding platforms now spreads this ability across the broader engineering team.

A project engineer can create the first version of a script using an AI assistant, refine it, and then share it through an ECP. This reduces waiting time, increases reuse, and captures knowledge within the project.

The role of the Computational Designer becomes more strategic, defining standards, validating outputs, and managing shared code libraries. Coding becomes part of daily engineering rather than a separate specialisation.

5. Keeping engineering rules central

AI can generate code, but it can’t decide what is right! That judgement still belongs to programmers

Every automation must remain true in design codes, safety factors, and sound engineering reasoning.

Modern ECPs make it easy to encode these rules into reusable components, verified once, applied many times. Over time, this builds a living library of proven engineering logic. The system handles repetition, but the principles remain human defined and professionally checked.

This approach keeps responsibility where it belongs while improving consistency and speed.

6. The financial promise of Computational Design (CoDe)

CoDe has a direct financial impact to any AEC business.  Global studies and project data show clear, repeatable savings when design automation and data-driven methods are applied.

7. Typical project gains

• 60–80% faster design coordination and data workflows.
• 40–60% faster quantity estimation and model updates.
• 10–20% lower material waste and rework.
• Up to 70% fewer manual adjustments in complex geometry changes.

8. Return on Investment

Typical ROI from Computational Design ranges between 10% and 13%, depending on scope and integration.

Savings by project phase often follow:

• Planning and Studies: 25–30%
• Design and Delivery: 25–30%
• Operations and Maintenance: 1–5%

9. Value-based delivery models

Traditional hourly billing does not reflect these gains. When work is billed by time, faster delivery reduces revenue. Under lump sum or value-based contracts, the time saved turns directly into margin and additional project capacity.

This changes the economics of engineering. Efficiency becomes an advantage to delivering more projects with the same resources and improving overall profitability.

10. How to do more, with the same amount of Engineers

Statistics don’t lie.  The table below shows that the institutions producing our awesome engineering talent are not producing the numbers that we need to sustain our infrastructure needs.  The question is then; how do we do more with the same amount of talent?

The answer is simple, maximise the use of technology, especially the ones that produce exponential output.

11. The shift

The evolution of Computational Design is taking place right now. Teams are moving from visual scripts and manual debugging to AI-assisted coding that captures design intent in structured, reusable logic.

Engineering Coding Platforms and AI assistants make it possible to generate and deploy code rapidly, keeping engineers in control while increasing throughput.

This is what it means to super charge Computational Design:

• Engineers and programmers define the rules.
• AI handles the coding and repetition.
• Projects run faster, cleaner, and with consistent quality.

Engineering staff won’t be replaced but we must give our engineering teams the time and tools to solve bigger problems with better precision and less friction.

12. The choice ahead

We now face a very clear choice i.e. stay with the traditional, linear ways of working, sequential modelling, disconnected tools, and slow iteration, or adopt Engineering Coding Platforms that combine engineering logic with AI speed.

The risk of staying with the old approach is not just slower delivery, it’s losing competitiveness. Two of the world’s largest engineering firms have already moved firmly in this direction, building ECP-based workflows and AI-driven design frameworks across their global operations. Those who do not adapt will soon find themselves explaining why their delivery times, margins, and design quality lag behind the leaders.

Computational Design has matured. The tools are ready. The question now is whether we are ready to keep pace….to me the answer is simple…