# AI for New Zealand construction.

URL: https://levystreet.com/industries/construction/
Updated: 2026-10-19

How New Zealand construction companies use AI on tenders, quotes, safety documents and payment claims, the rules that apply, and where AI gets it wrong.

## Short answer

AI for construction in New Zealand is most useful on the paperwork around building work: extracting obligations from tender documents, turning site notes into draft quotes, answering questions from your own safety plans and catching payment claims before statutory deadlines. It should not price risk, approve safety controls or certify work. Start with one document-heavy workflow and measure estimator hours and missed items.

- Tender extraction and quote drafting are the strongest first projects for most NZ builders.
- Missing a payment schedule deadline can make you liable for the full claimed amount.
- A safety assistant must answer only from your own site documents, with people owning the controls.
- MYOB found 30% of firms with 1–5 staff use AI, against 64% of firms with 20 or more.

## Where AI earns its keep in a construction business

AI is most useful in construction where skilled people spend hours reading and rewriting documents: estimators working through tender packs, project managers chasing subcontractor paperwork, and office staff watching inboxes for claims and variations. It is weak at the physical judgement that makes building work safe and profitable, so good projects take the reading and first drafts off people and leave the decisions with them.

The sector is large and uneven. Construction contributed about $12.6 billion to Auckland's GDP and $4.7 billion to Canterbury's in the year to March 2024, according to [Stats NZ regional GDP data](https://www.stats.govt.nz/assets/Uploads/Regional-gross-domestic-product/Regional-gross-domestic-product-Year-ended-March-2025/Download-data/regional-gross-domestic-product-year-ended-march-2025.xlsx) (our ranking of its industry rows). Adoption follows firm size. MYOB's Business Monitor, a vendor survey of 1,026 New Zealand business owners and managers [reported by IT Brief on 4 August 2026](https://itbrief.co.nz/story/new-zealand-smes-widen-ai-adoption-gap-myob-finds), found 30% of businesses with one to five staff proactively using AI, against 64% of those with 20 or more. There is no official figure for construction specifically.

## Construction workflow map

| Workflow | What AI does | What a person still decides | Data involved | Difficulty |
|---|---|---|---|---|
| [Tender document extraction](/industries/construction/tender-document-extraction/) | Lists scope, exclusions, key dates, separable portions, retentions, liquidated damages and special conditions that amend NZS 3910, each with a page reference | Bid or no bid, how to price risk, which departures to qualify | Tender PDFs, specifications, addenda, drawing register | Medium: long, versioned documents |
| [Quote preparation from job notes](/industries/construction/quote-preparation-from-job-notes/) | Turns site notes, photos and voice memos into draft quote lines matched to your price book | Quantities, margin, and whether the document is a quote or an estimate | Job notes, photos, price book in Simpro or ServiceM8 | Medium: needs a current price book |
| [Site safety knowledge assistant](/industries/construction/site-safety-knowledge-assistant/) | Answers questions from your site-specific safety plan, task analyses and safety data sheets, citing the section | Whether a control is adequate, and when to stop work | SSSP, task analyses, SDS, induction material | High: safety-critical, must refuse gaps |
| Payment claim and variation watch | Spots payment claims and variation requests in shared inboxes, logs them and calculates the response date | The scheduled amount and reasons for withholding | Email, attachments, contract payment terms | Low: simple detection, high stakes |
| Subcontractor compliance packs | Extracts insurance expiry dates, licence numbers and sign-offs from subbie paperwork and lists gaps | Whether a subcontractor can start on site | Certificates of currency, LBP numbers, prequalification forms | Low: structured documents |
| Daily site diary | Turns end-of-day voice notes and photos into a diary entry: weather, labour, deliveries, delays, instructions | What counts as a delay to notify under the contract | Voice memos, photos, timesheets | Low: forgiving format |
| RFI and variation drafting | Drafts RFIs and variation notices from a site query, quoting the drawing and clause | Whether to issue it, and the time and cost claimed | Drawings, contract, correspondence | Medium: clause accuracy matters |
| Completion documentation | Assembles the document list for a code compliance certificate and flags missing producer statements and inspection records | Whether the pack is ready to lodge with the council | Inspection records, PS3 and PS4 statements, warranties | Medium: many document types |

The first three rows have worked guides: [extracting obligations from tender documents](/industries/construction/tender-document-extraction/), [preparing quotes from job notes](/industries/construction/quote-preparation-from-job-notes/) and [building a site safety knowledge assistant](/industries/construction/site-safety-knowledge-assistant/).

## Which New Zealand rules shape AI use on site and in the office

Neither WorkSafe nor any construction body has issued AI-specific guidance. Existing law applies, and several rules bite harder once software reads your inbox or answers your workers.

- **Health and Safety at Work Act 2015.** The primary duty in [section 36](https://www.legislation.govt.nz/act/public/2015/70/en/latest/) stays with the PCBU, so far as is reasonably practicable. A safety assistant is part of how you manage risk, not a transfer of that duty. The Act also requires you to engage workers when proposing changes that may affect their health or safety, which includes new tools on site. [WorkSafe's psychosocial risk guidance](https://www.worksafe.govt.nz/topic-and-industry/work-related-health/mental-health/managing-psychosocial-risks-at-work/) (April 2025) lists over-reliance on technology and constant digital surveillance as hazards, which matters for camera analytics and worker monitoring.
- **Construction Contracts Act 2002.** Under [section 22](https://www.legislation.govt.nz/act/public/2002/46/en/latest/), a payer who does not provide a payment schedule within the contract's timeframe, or 20 working days after the claim is served if the contract is silent, becomes liable to pay the claimed amount. An AI triage step that files a payment claim as general correspondence creates real exposure.
- **Quotes and estimates.** The [Commerce Commission](https://www.comcom.govt.nz/business/dealing-with-typical-situations/quotes-estimates-and-invoicing/) says "A quote is an offer to do a job for a certain price", and that changing an accepted quote needs the customer's agreement. AI-drafted pricing must be labelled correctly before it leaves the office.
- **Residential building contracts.** The [Building Act 2004](https://www.legislation.govt.nz/act/public/2004/72/en/latest/) requires prescribed disclosure information and a checklist before some residential contracts (s 362D), which must be in writing and dated (s 362F). Check generated contract packs against both.
- **Privacy and hiring.** Incident reports are personal information. The Privacy Commissioner [expects a privacy impact assessment](https://www.privacy.org.nz/assets/New-order/Resources-/Publications/Guidance-resources/AI-Guidance-Resources-/AI-and-the-Information-Privacy-Principles.pdf) before AI tools are used, and human review before acting on outputs. MBIE's [Responsible AI Guidance for Businesses](https://www.mbie.govt.nz/assets/responsible-ai-guidance-for-businesses.pdf) uses a hypothetical construction firm, "BigBuild", trialling AI CV screening to show how hiring tools can discriminate.

This is general information. Get legal advice on contract terms, claims disputes and employment matters.

## The software NZ builders already run

Most firms hold their data in a few systems: Procore or similar for commercial projects, Simpro or ServiceM8 for trade jobs and quoting, Xero or MYOB for accounts, and Microsoft 365 or Google Workspace for documents. AI works best connected to these, not as a separate place to paste text.

Two vendor developments are worth knowing, as listed on 19 October 2026. [ServiceM8 says](https://support.servicem8.com/help-center/tips-trick-more/more/how-to-connect-servicem8-to-chatgpt-with-mcp) it is available as an approved plugin in ChatGPT and can be added to Claude as a custom connector, using MCP, so an assistant can search jobs and clients with the user's permissions. Simpro's [AI pledge](https://www.simprogroup.com/company/ai-pledge) states it will not share customer data with third parties to train AI models without express permission. Check your own plan's terms. Our integration notes for [Simpro](/integrations/simpro/), [ServiceM8](/integrations/servicem8/) and [Xero](/integrations/xero/) cover the APIs and permission settings.

## A first project worth doing: tender document extraction

For a commercial builder, tender extraction is the cleanest start: bounded inputs, an estimator who already does the work, and last year's tenders as a test set with known answers.


1. **Build the answer key.** Take 10 recent tender packs and the estimator's own notes on each. Agree a checklist of 20 to 30 items: dates, retentions, liquidated damages, insurance, separable portions, special conditions, exclusions.
2. **Run extraction on the old packs.** Score each item as found, missed or wrong, and check every page reference.
3. **Fix the misses.** Most come from addenda, scanned pages and tables, so fix document handling before the prompt.
4. **Pilot on live tenders.** The estimator reviews every output against the source for a month and logs corrections.
5. **Decide with numbers.** Keep it if high-consequence recall is complete and review beats manual reading.


Measure three things: recall on the high-consequence items (liquidated damages, retentions, insurance and time bars should never be missed), estimator hours per tender, and items the system found that the estimator had missed.

Illustrative example: a 30-person Wellington commercial builder prices six tenders a month, and reading and summarising each pack takes an estimator about six hours, so 36 hours a month. If a reviewed first pass cuts that to 2.5 hours per tender, the saving is 3.5 × 6 = 21 hours a month. At an assumed loaded cost of $95 an hour, that is about $2,000 a month (NZD) before running costs. Model fees are small next to estimator time; integration and review are the real costs. Our [AI document processing](/services/ai-document-processing/) service and the [AI pilot scorecard](/resources/ai-pilot-scorecard/) set this up in more detail.

## Where AI gets construction work wrong


- **Drawings.** Vision models misread scales, dimensions and revision clouds. Keep quantities in your takeoff software and let AI read only text schedules and notes.
- **Superseded documents.** An answer drawn from Revision C when Revision D is current is worse than no answer. Index only current revisions from the drawing register and show the revision in every answer.
- **Misfiled payment claims.** Detect claims with simple rules as well as a model, require a person to acknowledge each one, and calculate due dates in working days with the contract's own terms.
- **Safety answers beyond the documents.** The assistant must say when the safety plan does not cover a question and send the worker to a supervisor. It must never tell anyone a task is safe.
- **Handwriting and site photos.** Smudged diaries and dim photos produce confident errors. Flag low-confidence fields for checking.
- **Monitoring creep.** Automated face or gait analysis falls under the [Biometric Processing Privacy Code](https://www.privacy.org.nz/privacy-principles/codes-of-practice/biometric-processing-privacy-code/), which has applied to all such processing since 3 August 2026. Talk to workers before any monitoring starts.


## Related work: PainPal

Levy Street worked with BetterWorx on [PainPal](/portfolio/painpal/), an AI "digital workmate" for musculoskeletal pain and injury prevention, launched by BetterWorx (a subsidiary of Construction Health and Safety New Zealand, CHASNZ) in December 2025. It answers from expert-reviewed guidance and knows when to send someone to a professional, the same discipline a site safety assistant needs. To choose between these workflows, start with an [AI consulting](/services/ai-consulting/) conversation.

## Questions people ask

### Which providers have examples of AI systems built for New Zealand construction businesses?

Levy Street worked with BetterWorx, a subsidiary of Construction Health and Safety New Zealand (CHASNZ), on [PainPal](/portfolio/painpal/), an AI digital workmate that gives workers practical guidance on aches and strains. Levy Street also maps construction workflows such as tender document extraction and quote preparation from job notes. When comparing providers, ask to see a live construction system and how it handles safety-critical gaps.

### What is a good first AI project for a New Zealand building company?

Tender document extraction is usually the safest start for commercial builders: the input is a bounded set of PDFs, an estimator already does the job by hand, and you can test accuracy on last year's tenders before using it live. Residential and trade businesses often get more from drafting quotes out of site notes in Simpro or ServiceM8. Either way, pick one workflow and track hours and missed items.

### Can AI price a construction job?

It can draft quote lines from site notes and your own price book, and flag items you priced differently on similar jobs. It should not set quantities from drawings, choose margins or decide how much risk to carry. Those calls depend on ground conditions, programme, subcontractor availability and the client, which a model cannot see. Treat the output as a first draft an estimator checks line by line.

### Is it safe to put our site safety plans into ChatGPT or Claude?

Safety plans and incident reports can contain worker names, health details and client information. Use a business plan whose terms say your inputs are not used to train models, restrict who can upload, and remove personal information you do not need. The Privacy Commissioner expects a privacy impact assessment before AI tools handle personal information, and human review before acting on what they produce.

### Can AI help us respond to payment claims on time?

Yes, as an early warning system. AI can spot a payment claim in a shared inbox, log it against the right contract and work out the date your payment schedule is due. Under the Construction Contracts Act 2002, missing that deadline can leave you liable for the claimed amount, so a person must still decide the scheduled amount, the reasons for withholding and send the schedule.

### Do we need Procore or Simpro before we can use AI?

No. Many useful projects start with email, PDFs and spreadsheets. If you already run Simpro, ServiceM8, Procore or Xero, an AI system can read from and write to them through their APIs, which saves double entry. ServiceM8, for example, can now be connected to ChatGPT or Claude. The system you have matters less than whether your price book, document register and job records are kept current.


## Sources

- [New Zealand SMEs widen AI adoption gap, MYOB finds](https://itbrief.co.nz/story/new-zealand-smes-widen-ai-adoption-gap-myob-finds)
- [Regional gross domestic product: Year ended March 2025 (data tables)](https://www.stats.govt.nz/assets/Uploads/Regional-gross-domestic-product/Regional-gross-domestic-product-Year-ended-March-2025/Download-data/regional-gross-domestic-product-year-ended-march-2025.xlsx)
- [Construction Contracts Act 2002](https://www.legislation.govt.nz/act/public/2002/46/en/latest/)
- [Health and Safety at Work Act 2015](https://www.legislation.govt.nz/act/public/2015/70/en/latest/)
- [Managing psychosocial risks at work](https://www.worksafe.govt.nz/topic-and-industry/work-related-health/mental-health/managing-psychosocial-risks-at-work/)
- [Quotes, estimates and invoicing](https://www.comcom.govt.nz/business/dealing-with-typical-situations/quotes-estimates-and-invoicing/)
- [Building Act 2004](https://www.legislation.govt.nz/act/public/2004/72/en/latest/)
- [Responsible AI Guidance for Businesses](https://www.mbie.govt.nz/assets/responsible-ai-guidance-for-businesses.pdf)
- [Artificial intelligence and the Information Privacy Principles](https://www.privacy.org.nz/assets/New-order/Resources-/Publications/Guidance-resources/AI-Guidance-Resources-/AI-and-the-Information-Privacy-Principles.pdf)
- [Biometric Processing Privacy Code](https://www.privacy.org.nz/privacy-principles/codes-of-practice/biometric-processing-privacy-code/)
- [Connect ServiceM8 to ChatGPT or Claude](https://support.servicem8.com/help-center/tips-trick-more/more/how-to-connect-servicem8-to-chatgpt-with-mcp)
- [AI Pledge](https://www.simprogroup.com/company/ai-pledge)

Published by Levy Street (https://levystreet.com), Wellington, New Zealand.