AI for Construction Contract Review — A Practicing QS Professional's Guide
The Sentence That Could Have Cost Millions
I want to start with a real example because it illustrates better than any theoretical explanation why construction contract review requires professional expertise that no AI tool can replicate.
On an expressway project tender, the main contract conditions stated clearly that excavated material from cut areas should be used for filling at the embankment locations and that payment would be made for both excavation and embankment filling as standard pay items. On the surface, this looked straightforward: excavate, reuse, fill, get paid.
Most QS professionals reviewing this contract would have priced it exactly that way, no disposal cost, full material reuse assumed, clean and simple.
But buried in the Technical Specification, under the material testing requirements section, was this sentence:
"Materials excavated from the cut area with a Plasticity Index (PI) greater than 15, or containing organic or shale fragments, shall be deemed unsuitable for structural embankment fill and must be disposed of at the Contractor's sole expense."
Read that carefully. At the Contractor's sole expense.
The main contract said reuse everything. The specification said if the material fails the PI test, dispose of it. At your own cost. No separate pay item. No recovery mechanism.
On an expressway project, excavation volumes can run into hundreds of thousands of cubic metres. If a significant proportion of that material has a Plasticity Index above 15, which is not uncommon in certain geological conditions, the disposal cost can reach millions. And the Contractor would have no contractual basis to recover a single dollar of it, because they agreed to a contract that clearly stated this obligation in the specification.
The solution was straightforward: obtain the geotechnical report, review the soil classification data for the cut areas, estimate the proportion of material likely to exceed PI 15, calculate the disposal volume and cost, and incorporate it into the tender price.
That one specification sentence was found only because I read every section of the material testing requirements, which changed our tender pricing significantly.
This is what professional contract review looks like. And this is why the question of how AI fits into that process deserves a careful, honest answer.
📚 Related Articles You May Find Useful:
The Core Challenge Construction Contract Document Complexity
The fundamental challenge in construction contract review is not the complexity of any single document; it is the volume and interconnected nature of the complete contract document suite, and the way critical obligations can be distributed across multiple documents that most people never read in full.
The PI specification example above illustrates this perfectly. The main contract said one thing. The specification said something different, and the specification governed material suitability. A QS who only read the main contract conditions and assumed material reuse would have missed a potentially catastrophic cost obligation.
| Contract Document | Primary Purpose | Key QS Relevance |
|---|---|---|
| Conditions of Contract | Legal and contractual framework | Variation, claims, payment, dispute provisions |
| Technical Specifications | Technical requirements and standards | Material standards, testing requirements hidden cost obligations |
| Drawings | Design intent and scope definition | Quantity take-off basis, scope boundaries, General Notes |
| Bill of Quantities | Measured works and pricing schedule | Directly linked to Specification and Drawings |
| Method Statements | Construction methodology | Productivity assumptions, resource planning |
| Employer Requirements | Client-specific project requirements | Scope inclusions and exclusions |
| Tender Clarifications | Pre-contract clarifications and agreements | Scope adjustments, agreed deviations |
Each document is linked to the others. The BOQ references the Specification. The Specification references the Drawings. The Conditions of Contract govern how all documents interact. And cost-significant obligations can be buried in any of them.
How I Use AI for Construction Contract Document Review
My application of AI tools in contract review is specific, controlled, and professionally bounded.
On my current large-scale infrastructure project, the technical specifications run to hundreds of pages across multiple work packages. Retaining all relevant technical provisions across those packages through reading alone is not practically achievable, and that creates risk. The kind of risk that produces situations like the PI specification example above.
The primary AI application I have found genuinely valuable is using ChatGPT to summarize large specification sections, giving me a structured overview of the key technical requirements, material standards, testing obligations, and cost-significant provisions in a fraction of the time that manual reading would require.
My Methodology ChatGPT Project Folder for Specification Review
| Stage | Action | Purpose |
|---|---|---|
| Stage 1 | Create a dedicated Project folder in ChatGPT | Isolates project data within a controlled workspace |
| Stage 2 | Upload the relevant specification document | Provides ChatGPT with the source document for reference |
| Stage 3 | Configure settings to "Project only" | Ensures answers are based solely on uploaded data, not general training data |
| Stage 4 | Submit specific summary requests | Extracts relevant technical provisions efficiently |
| Stage 5 | Request page number references with all responses | Enables independent verification against source document |
| Stage 6 | Independently verify against original document | Confirms accuracy before professional application |
The page number reference technique is critical. By instructing ChatGPT to provide the relevant page or section reference for each summary point, I can locate and verify the original provision directly rather than relying solely on the AI summary.
And here is the professional discipline that must always apply: AI summarization is useful for getting an overview quickly. But the detailed professional review looking for hidden obligations like the PI specification example still requires a qualified QS reading the full document carefully. AI gives you a map. The professional still needs to walk the terrain.
The Five Critical Contract Clauses Every QS Must Prioritise
Regardless of the contract form FIDIC, NEC, JCT, or bespoke, these five areas carry the greatest commercial significance for the Quantity Surveyor and must be reviewed with particular care before project commencement.
1. Variation Clause
The variation clause defines how changes to the original scope are identified, instructed, and valued. Key provisions to understand: what constitutes a variation, who has authority to instruct it, the valuation methodology, the process for agreeing values, and time limits for submitting variation claims.
2. Contract Price and Payment Terms
Whether the contract is lump sum, remeasurable, or cost-reimbursable fundamentally affects how commercial risk is distributed. Payment application process, certification timescales, retention provisions, and advance payment terms must all be clearly understood before project commencement.
3. Measurement and Evaluation
The applicable method of measurement, how permanent works are certified, treatment of provisional quantities, daywork provisions, and dispute resolution for measurement disagreements all must be identified and understood. And as the PI example demonstrates, the measurement basis described in the BOQ preambles must be cross-referenced against the technical specification to identify hidden inclusions and exclusions.
4. Claims Procedures and Time Bars (EOT)
💼 From My Current Project FIDIC Sub-Clause 20.1:
My current large-scale infrastructure project operates under the FIDIC Pink Book Conditions of Contract for Construction MDB Harmonised Edition (June 2010). Under Sub-Clause 20.1 Contractor's Claims, the contract states: "The notice shall be given as soon as practicable, and not later than 28 days after the Contractor became aware, or should have become aware, of the event or circumstance."
The consequence of failing to comply is unambiguous — if notice is not given within 28 days, the Time for Completion shall not be extended, the Contractor shall not be entitled to additional payment, and the Employer is discharged from all liability. On a large-scale infrastructure project, the financial implications of missing a single notice deadline can be substantial. This is why claims procedure management is a human professional responsibility; it cannot be delegated to an AI tool.
5. Priority of Documents
The priority of documents clause establishes which document governs in the event of conflict or ambiguity. This is directly relevant to situations like the PI specification example, where the main contract and the technical specification appeared to say different things about material reuse. Understanding document hierarchy tells you which provision governs and that determines your commercial position.
Summary Critical Clauses for QS Review
| Priority | Clause | Why Critical for QS | Risk if Missed |
|---|---|---|---|
| 1 | Variation Clause | Governs scope change valuation | Unrecovered variation costs |
| 2 | Contract Price & Payment Terms | Defines payment structure and process | Cash flow issues, missed entitlements |
| 3 | Measurement & Evaluation | Governs how work is valued — and what is included | Hidden cost obligations missed at tender stage |
| 4 | Claims Procedures & Time Bars (EOT) | Defines notice requirements and deadlines | Complete loss of valid claims |
| 5 | Priority of Documents | Establishes document hierarchy | Incorrect resolution of conflicts between documents |
Critical Professional Limitations: What AI Cannot Do in Contract Review
AI tools are not developed for commercial judgment.
Going back to the PI specification example could AI have found that sentence? Possibly, if asked to summarize the material testing section specifically. But AI cannot tell you what that sentence means commercially. It cannot tell you what proportion of your excavated material is likely to exceed PI 15 based on the geotechnical report. It cannot calculate the disposal cost exposure. And it cannot make the commercial judgment call about how much contingency to include in your tender price.
Those are human professional functions. They require construction knowledge, site experience, and commercial judgment that no AI system can replicate.
1. Risk Allocation Assessment
Identifying where contractual risk sits between Employer, Contractor, and Engineer requires professional judgment built on construction experience and understanding of commercial implications. The PI specification sentence transferred a potentially enormous disposal cost risk entirely to the Contractor. Identifying and pricing that risk was a human professional function.
2. Commercial Judgment
Determining the commercial significance of a clause and advising on pricing strategy based on contract terms requires professional expertise. AI can present the clause. It cannot assess what it means for your specific tender price.
3. Project-Specific Contextual Analysis
Every construction project has unique geological conditions, site constraints, and commercial context. AI has no knowledge of these factors and cannot apply project-specific judgment to contract analysis.
Data Protection: A Non-Negotiable Professional Obligation
Construction contracts contain highly sensitive commercial data, contract values, pricing structures, commercial terms, and confidential client information. My professional advice is unequivocal:
Never upload sensitive or confidential contract data to third-party public AI platforms.
This position is aligned with RICS guidance on the Responsible Use of Artificial Intelligence in Surveying Practice. QS professionals must ensure that confidential contract data is not uploaded to public AI platforms, company data protection policies are observed, client confidentiality agreements are respected, and applicable data protection legislation, including GDPR, is complied with.
Practical Guidelines AI-Assisted Contract Review
| Activity | AI Suitable? | Professional Requirement |
|---|---|---|
| Summarizing large specification sections | Yes - with verification | Verify all summaries against source document |
| Locating specific clause references | Yes - with verification | Confirm page/clause reference independently |
| Identifying key clause headings | Yes- with verification | Professional review of identified clauses required |
| Risk allocation assessment | No - Human only | Qualified QS professional judgment required |
| Commercial judgment on contract terms | No - Human only | Qualified QS professional judgment required |
| Claims strategy and notice management | No - Human only | Qualified QS professional judgment required |
| Uploading confidential contract data | No- Never | RICS guidance and data protection obligations apply |
Conclusion
The PI specification example I started with is not an unusual situation. Hidden cost obligations buried in technical specifications, while the main contract appears to say something different, are a standard feature of complex construction contracts. Finding them requires a qualified professional reading every section of every document with commercial awareness.
AI tools can help navigate those documents faster. They can summarize large sections efficiently, locate clause references quickly, and reduce the time required for initial document orientation. On a large infrastructure project with hundreds of pages of specification content, that time saving is real and valuable.
But AI cannot read between the lines. It cannot cross-reference a material testing clause against an excavation payment item and calculate your disposal cost exposure. It cannot make the commercial judgment call that protects your tender price.
Use AI to navigate the documents faster. Use your professional knowledge to understand what they actually mean.
💬 Found This Useful?
If this article helped you — please consider:
- ⭐ Sharing it on LinkedIn — the PI specification example could save a colleague significant money
- 💬 Leaving a comment below — have you found a hidden cost trap in a contract document? Share your experience!
- 📖 Reading our related articles — links below
What is the most dangerous hidden obligation you have found in a construction contract? Share it in the comments!
About The Author
A practicing Quantity Surveyor with extensive experience in large-scale infrastructure projects under FIDIC conditions.
Qualifications:
- BSc (Hons) Quantity Surveying
- MCIOB — Chartered Institute of Building
- MQSI — Quantity Surveyor Institute Member
This blog shares real professional experience from live construction projects — not theoretical content.
📚 Related Articles You May Find Useful:
Frequently Asked Questions
What is the Plasticity Index and why does it matter in construction contracts?
Which AI tool is most suitable for construction contract document review?
What is the most critical contract clause for a QS to review?
Does RICS provide guidance on using AI in quantity surveying practice?
Comments
Post a Comment