AI for System Design: A Guide for Australian Software Engineers (2026)

System design is one of the most intellectually demanding parts of software engineering. It requires understanding requirements, evaluating trade-offs, anticipating failure modes, and making decisions that will constrain the system for years. AI tools are changing how software engineers approach this work — not by making the decisions, but by making the thinking process faster and more rigorous.

AI as a Design Partner

The most valuable use of AI in system design is as a thinking partner. When you're working through a design problem, AI can:

• Challenge your assumptions: "What happens if the message queue goes down?" "How does this scale to ten times the current load?" "What's the failure mode if the third-party API is unavailable?"
• Suggest alternatives: "What are the trade-offs between a monolith and microservices for this use case?" "What are the alternatives to a relational database here?"
• Identify gaps: "What authentication and authorisation approach are you using?" "How are you handling data consistency across these services?"
• Explain patterns: "What is the saga pattern and when is it appropriate?" "How does event sourcing work and what are its trade-offs?"

This is different from asking AI to design the system for you. The design decisions require understanding your specific context — your team's skills, your organisation's constraints, your users' needs — that AI doesn't have. But AI can make your thinking more rigorous by asking the questions a senior engineer would ask.

Architecture Decision Records

Architecture Decision Records (ADRs) are a practice for documenting significant architectural decisions — what was decided, why, and what alternatives were considered. They're valuable for onboarding new team members, understanding why the system is the way it is, and avoiding re-litigating past decisions.

AI makes writing ADRs significantly faster. The structure of an ADR is consistent:

• Context: what is the situation that requires a decision?
• Decision: what was decided?
• Rationale: why was this decision made?
• Alternatives considered: what other options were evaluated and why were they rejected?
• Consequences: what are the implications of this decision?

Describe your decision to Claude or ChatGPT and ask it to draft an ADR. The output will need editing — AI doesn't know your specific context — but it's a much faster starting point than writing from scratch.

For Australian software engineers working in regulated industries (financial services, healthcare, government), ADRs are particularly valuable for demonstrating that architectural decisions were made thoughtfully and with appropriate consideration of security and compliance requirements.

Scalability and Performance Analysis

AI can help analyse system designs for scalability and performance issues before they become production problems.

Load analysis: Describe your system's architecture and expected load to an AI and ask it to identify potential bottlenecks. "We have a REST API with a PostgreSQL database. We expect five hundred requests per second at peak. Where are the likely bottlenecks?" AI can identify common patterns — database connection pool exhaustion, N+1 query problems, synchronous calls to slow external services — that are worth investigating.

Capacity planning: AI can help with back-of-envelope calculations for capacity planning. "If each user generates ten database writes per day and we have one hundred thousand users, what's our write throughput?" These calculations are straightforward but easy to get wrong; AI can check your arithmetic and identify factors you might have missed.

Caching strategy: AI can help design caching strategies — what to cache, where to cache it, how to handle cache invalidation. Cache invalidation is famously one of the hard problems in computer science; AI can help you think through the edge cases.

Designing for Australian Compliance Requirements

Australian software engineers working on systems that handle personal data, financial information, or health records need to design with compliance in mind from the start. Retrofitting compliance is expensive; designing for it upfront is much cheaper.

Privacy Act compliance: The Privacy Act 1988 and the Australian Privacy Principles (APPs) impose requirements on how personal information is collected, stored, used, and disclosed. AI can help you identify which APPs apply to your system and what design decisions they require.

APRA requirements: For systems in the financial services sector, APRA's Prudential Standard CPS 234 on information security imposes specific requirements on system design. AI can help you understand these requirements and identify design decisions that address them.

Health data: The My Health Records Act and state health legislation impose requirements on systems handling health data. AI can help identify the relevant requirements and their design implications.

Practical approach: Describe your system and its data handling to Claude and ask: "What Australian regulatory requirements apply to this system and what design decisions do they require?" The output is a starting point for a more detailed compliance review, not a substitute for legal advice.

API Design

API design is a significant part of system design for most modern software systems. AI tools are useful for several aspects of API design.

OpenAPI specification generation: Describe your API's resources and operations to an AI and ask it to generate an OpenAPI specification. This is faster than writing the specification from scratch and produces a consistent, well-structured document.

API design review: Paste your API design into Claude and ask it to review it for consistency, completeness, and adherence to REST principles. AI can identify inconsistencies in naming conventions, missing error responses, and other common API design issues.

Versioning strategy: AI can help you think through API versioning strategies — URL versioning, header versioning, content negotiation — and their trade-offs for your specific situation.

Breaking change analysis: When you're changing an existing API, AI can help identify which changes are breaking and which are backwards-compatible. This is particularly useful for public APIs where breaking changes affect external consumers.

Design Documentation

System design documentation is often neglected because it's time-consuming to write and quickly becomes outdated. AI tools make documentation faster to write and easier to maintain.

Architecture diagrams: While AI can't generate diagrams directly, it can generate the text descriptions and component lists that are the basis for diagrams. Tools like Mermaid and PlantUML allow you to generate diagrams from text descriptions; AI can help write those descriptions.

Runbooks: AI can help write operational runbooks — step-by-step guides for common operational tasks like deployments, rollbacks, and incident response. Describe the procedure to AI and ask it to format it as a runbook.

Onboarding documentation: AI can help write onboarding documentation for new team members — system overview, key concepts, development environment setup, and common tasks.

Building Better Systems

The software engineers getting the most value from AI in system design are those who use it to make their thinking more rigorous, not to shortcut the thinking process. AI is a tool for exploring the design space more thoroughly, documenting decisions more completely, and identifying issues earlier — not a substitute for the engineering judgement that good system design requires.

The best system designs come from engineers who deeply understand the problem, the constraints, and the trade-offs. AI helps you explore those trade-offs more thoroughly and document your reasoning more clearly.