Artificial Intelligence (AI) is increasingly part of the engineer’s toolkit. Not because it replaces engineers, but because it helps us make better decisions faster, especially when working under pressure, with complex regulations, or constrained resources.

This guide introduces the principles and practices of using general-purpose AI tools (such as ChatGPT, Bing, Claude, Bard and ChatPDF) in a responsible, professional and compliant way across engineering disciplines in Aotearoa. It also includes AI tool examples, which illustrate how engineers are applying AI to real-world tasks like structural load checks.

On this page:

The role of AI in engineering practice

AI is best seen as an assistant, not a decision-maker.

It can:

  • Summarise long technical reports, standards, or site logs
  • Suggest corrective actions based on incident data or historical failures
  • Perform repetitive calculations, like factored loads or infiltration rates
  • Draft risk reports, safety summaries, or compliance statements
  • Check for overlooked load combinations or code clauses
  • Interpret council guidance or Building Code clauses when uploaded

AI can't:

  • Make engineering judgments
  • Sign off work
  • Automatically ensure code compliance
  • Replace site investigation, physical inspection, or peer review

Used appropriately, AI tools help engineers focus on complex reasoning and creative problem-solving, while reducing the time spent on mechanical tasks like document searching or first-draft preparation.

Guidance for integrating AI tools into your practice

  1. Start with clear intent and good prompts

AI works best when you are precise. Always:

  • Set the context (e.g. “You are reviewing a fire incident for New Zealand compliance.”)
  • Include relevant data or excerpts (e.g. your load combination sheet or percolation test results)
  • Be specific (e.g. “Check if this matches NZS 1170 ultimate limit state combos.”)
  • Ask for structure (e.g. “Provide in bullet points” or “Include clause references.”)

Good prompting turns AI from a generic chatbot into a capable assistant aligned to your task.

  1. Stay grounded in New Zealand codes and standards

AI can hallucinate or oversimplify. It may invent clauses or misquote data unless grounded in the actual document.

Use tools like ChatPDF, Claude, or Gemini to:

  • Upload New Zealand standards or Building Code excerpts
  • Ask for citations from within the actual document
  • Cross-check AI suggestions with the official wording

Always read the referenced clause yourself before accepting a recommendation.

3. Maintain professional oversight and ethical control

Treat AI like a junior colleague:

  • Review all outputs before use
  • Correct factual errors or oversights
  • Add contextual knowledge the AI might not have (e.g. site-specific risks, past failures, soil conditions)

Never:

  • Input confidential client data into public AI tools
  • Rely solely on AI for code compliance or safety assessments
  • Use AI to generate false, misleading, or fictitious content

Follow your company’s policies on data handling and seek client consent before using AI on project data.

4. Use AI to support, not skip, critical engineering steps

AI can support:

  • Drafting sections of technical reports
  • Generating checklists
  • Interpreting historical maintenance logs
  • Creating calculation templates

AI cannot replace:

  • Geotechnical judgement in soak pit siting
  • Manual review of incident trends for root cause analysis
  • Physical inspection of asset deterioration
  • Confirmation that control systems meet performance requirements

Engineers must verify all key outputs and remain liable for decisions under the Building Act, HSWA 2015, and Engineering New Zealand Code of Ethical Conduct.

Prompting frameworks

The following are three prompting frameworks that can be used for engineering applications:

  1. RTF (Role, Task, Format)

The RTF framework is a widely used and beginner-friendly structure for crafting effective prompts. It consists of:

  • Role: Defines who the AI should act as (e.g., "a nutritionist", "a historian").
  • Task: Specifies what the AI should do (e.g., "list benefits", "explain a concept").
  • Format: Indicates how the response should be structured (e.g., "bullet points", "table", "paragraph").

Use Cases:

  • Quick summaries
  • Lists of ideas or steps
  • Structured explanations

Examples: 

  • "As a structural engineer (role), outline 5 key considerations when designing a suspension bridge (task) in a numbered list format (format)."
  • "As a civil engineer, explain the steps involved in load analysis for a pedestrian bridge using bullet points."
  • "As a bridge design consultant, summarize the pros and cons of using cable-stayed vs. arch bridges in a table format."

For more information see visit.

  1. CREATE: Character, Request, Examples, Adjustments, Type, Extras

The CREATE framework is more advanced and comprehensive, ideal for nuanced or creative tasks. It includes:

  • Character: Assigns a persona or role to the AI (e.g., "a witty novelist", "a UX designer").
  • Request: Clearly defines the task.
  • Examples: Provides sample outputs to guide the AI.
  • Adjustments: Allows for refinements or iterations.
  • Type: Specifies the desired output format (e.g., story, list, dialogue).
  • Extras: Adds any additional context or constraints.

Use Cases:

  • Creative writing
  • Iterative content generation
  • Complex formatting or tone control

Example:

"As a senior structural engineer (Character), draft a conceptual design brief for a new highway bridge (Request), similar to the example provided below (Examples), but adapted for a coastal environment with high wind loads (Adjustments). Format it as a technical report summary (Type). Include sustainability considerations and material selection rationale (Extras)."

For more information visit.

  1. CRAFT: Context, Role, Action, Format, Tone

The CRAFT framework is a practical and structured approach to prompt writing, particularly useful for professionals aiming to generate clear, relevant and fit-for-purpose outputs. It works well for technical, analytical, and communication tasks where tone and precision matter.

It consists of:

  • Context: Provides background or situational detail (e.g. "You're assisting with a design review", "This is for a client-facing report").
  • Role: Defines who the AI should act as (e.g. "a senior engineer", "a technical editor").
  • Action: States the task clearly (e.g. "review this summary", "generate a risk register").
  • Format: Specifies how the output should be presented (e.g. table, checklist, email).
  • Tone: Indicates the desired style or formality (e.g. "professional", "plain English", "advice to a peer").

Use cases:

  • Drafting emails or reports
  • Reviewing technical content
  • Preparing stakeholder communications
  • Documenting assumptions or risks

Examples:

  • "Context: This is for a proposal to a local council. Role: You are a senior civil engineer. Action: Write a summary of the drainage strategy. Format: Use plain English in paragraph form. Tone: Professional but accessible."
  • "You're reviewing the structural design summary of a timber building (context). As a building consent authority reviewer (role), identify potential red flags or missing checks (action), and list them in bullet points (format) using a neutral and constructive tone (tone)."

For more information go here.

Capabilities across engineering disciplines

Discipline

Typical AI applications

Professional considerations and limits

Discipline Building Services / HVAC

Typical AI applications Analysing operational data, assisting in control logic development, estimating energy savings

Professional considerations and limits Requires manual validation of comfort outcomes and compliance with ventilation standards

Discipline Structural Engineering

Typical AI applications Drafting documentation, interpreting code clauses, checking combination logic

Professional considerations and limits Structural safety decisions must be verified against standards and engineering judgement

Discipline Civil / Stormwater

Typical AI applications Performing runoff and soakage calculations, summarising guidelines, sizing detention systems

Professional considerations and limits Final designs depend on site-specific testing, local rules, and peer review

Discipline Geotechnical Engineering

Typical AI applications Explaining soil behaviour concepts, generating draft summaries of site investigation logs

Professional considerations and limits Cannot interpret field or lab test results or replace geotechnical modelling

Discipline Infrastructure Asset Management

Typical AI applications Prioritising maintenance based on condition logs, forecasting lifecycle risk

Professional considerations and limits Must be calibrated to local conditions and reviewed against physical inspection records

Discipline Fire Safety

Typical AI applications Summarising incident logs, referencing regulatory clauses, drafting “lessons learned” content

Professional considerations and limits Fire scenarios must be assessed by qualified personnel; AI cannot determine causality

Discipline Environmental Engineering

Typical AI applications Summarising environmental assessments, suggesting monitoring strategies, interpreting rules

Professional considerations and limits Regulatory advice must be confirmed; environmental impacts require empirical evidence

Discipline Transportation / Roads

Typical AI applications Scheduling pavement renewals, analysing traffic volume data, summarising asset performance

Professional considerations and limits Designs and forecasts must consider geographic, safety, and human behavioural factors

Discipline Water and Wastewater

Typical AI applications Drafting O&M plans, summarising treatment system requirements, assisting with regulatory queries

Professional considerations and limits Operational design and compliance decisions must be based on expert review

Discipline Utilities (Electricity, Gas)

Typical AI applications Interpreting control system documentation, assisting with fault log reviews

Professional considerations and limits Safety-critical systems require extensive validation and testing procedures

Discipline Project and Construction Management

Typical AI applications Generating task checklists, tracking issue logs, producing status summaries

Professional considerations and limits Outputs must be aligned with site reality, team roles, and health and safety protocols

Discipline Engineering Education and Training

Typical AI applications Creating training content, generating quiz questions, simplifying technical concepts

Professional considerations and limits Content should be reviewed for technical accuracy and aligned with curriculum standards

Discipline Policy and Regulation Support

Typical AI applications Drafting regulatory summaries, comparing statutory requirements across jurisdictions

Professional considerations and limits Outputs must be checked for currency, authority, and legal compliance

Check out AI engineering tools for additional information on specific AI tools for your engineering discipline.

Building your company’s AI maturity

Start simple:

  • Use free tools like ChatGPT, Gemini, Bard, or Claude. Note: These tools are constantly being updated, and you need to ensure your data remains secure.
  • If you find a tool you prefer, build your experience and confidence in this tool
  • Trial AI on internal documents or non-client-facing workflows (e.g. summarising a previous report)
  • Develop internal prompting templates for repeatable tasks
  • Hold a team session on AI do’s and don’ts

Then grow:

  • Assign a staff member to stay across updates and new tools
  • Create a shared log of useful prompts or example successes
  • Trial ChatPDF or Claude with New Zealand standards and council guidelines
  • Always verify AI outputs – engineers keep liability, not the AI
  • Document where and how AI was used in project files (for audit and transparency)

If confidentiality is critical:

  • Consider open-source AI like LLaMA 2 or GPT4
  • Mask identifiable data in AI prompts
  • Use opt-out features to prevent training on your data

AI tools for discipline specific applications

The following guide demonstrates how structural engineers in New Zealand can use AI tools to assist with discipline specific applications.

Checking structural load combinations

This guide includes

  • Examples of AI-assisted infrastructure workflows
  • AI prompting strategies for accurate, New Zealand-specific results
  • A real-world example of using AI with detailed steps and references to applicable codes, standards, or regulatory sources
  • Good practice reminders for professional use

This guide is designed to be adapted. Start small, apply it to non-critical tasks, and build confidence in how AI can support, not replace, your engineering judgement.

More guides will be available at a later date.

Our use of AI

Content creation

Artificial Intelligence is a powerful tool that can help Engineering New Zealand Te Ao Rangahau in our work, including creating content.

We take a managed approach to how we use AI in content creation. We do so with the aim of enhancing our productivity and creativity, and allowing us to better serve our members, the engineering profession, and the public.

Read more

AI use policy
Help inform your own AI use policy by checking out our one.
View the policy

Contact us

Your thoughts and insights matter to us. Engage with us as we explore AI's role in engineering together.
Get in touch