How Quantity Surveyors Can Use AI to Prepare Robust Delay and Disruption Cost Claims
Delay claims are won or lost on the quality of your contemporaneous records — not the size of your programme float. By the time your client is in dispute, the paper trail you built during the works is the only thing standing between a successful recovery and a settlement that doesn’t cover your preliminaries. If you’re still manually reconciling daily reports, RFI logs, and cost data in a spreadsheet at midnight before a submission deadline, there’s a better way. AI for construction delay cost claims is changing how QS practitioners structure evidence, calculate prolongation heads, and produce claim narratives — and the tools are mature enough to use on live projects right now.
flowchart TD
A["Delay Event Occurs"] --> B{"Contemporaneous Records Exist?"}
B -->|No| C["Collect Available Evidence"]
B -->|Yes| D["Organize Records with AI"]
C --> D
D --> E["AI Calculates Prolongation Costs"]
E --> F["Generate Compliant Narrative"]
F --> G["Submit Robust Delay Claim"]
G --> H["Claim Success"]
1. Using AI to Collate Contemporaneous Records for Delay and Disruption Claims Construction AI Workflows
When you’re sitting in the site office at 5pm with three months of daily reports, a weather log, an RFI register, and a subcontractor correspondence folder, the first problem isn’t analysis — it’s ingestion. Getting all of that into a coherent structure before you can do anything useful with it is where most QS professionals lose days.
AI document analysis tools remove that bottleneck. ChatGPT-4o (from $20/month via ChatGPT Plus, or API access from approximately $0.005 per 1K tokens) can ingest uploaded PDFs and CSVs and identify delay events, critical path impacts, and employer risk events across hundreds of pages in minutes. Claude 3.5 Sonnet by Anthropic (free tier available; Pro from $20/month) handles large context windows well — useful when you’re feeding it a 200-page site diary alongside a programme baseline.
Here’s a practical step-by-step for getting your contemporaneous records into a usable claim structure:
Step 1: Export your RFI register and daily report log as CSVs — This creates machine-readable source data that any AI tool can cross-reference against your programme activities.
Step 2: Upload both files to Claude or ChatGPT-4o alongside your baseline programme (as a PDF export) — Ask the model to identify RFIs that correspond to critical path activities and flag any that remained unanswered beyond the contractual response period.
Step 3: Ask the AI to generate a delay event schedule — Include columns for event description, employer/contractor/neutral risk classification, date first notified, and associated programme activity ID.
Step 4: Review and validate the output against your site diaries — AI will occasionally misread dates or conflate two separate events. A 30-minute QS review catches these before they end up in a formal submission.
Step 5: Save the structured schedule as your claim annexure foundation — Every delay event in your narrative should trace back to a row in this register.
how to structure a delay event register
Try this prompt:
You are a construction claims analyst. I am attaching three documents: (1) an RFI register in CSV format for the [PROJECT NAME] project, (2) daily site reports from [START DATE] to [END DATE] in PDF, and (3) a baseline programme exported as PDF. Your task is to identify all employer-risk delay events that affected critical path activities. For each event, list: the RFI or instruction number, the date issued, the contractual response deadline, the date responded (if recorded), the delay in calendar days, and the affected programme activity. Present this as a table sorted by event date. Flag any events with a delay greater than 14 days.
2. How to Automate Prolongation Cost Calculation Using AI Quantity Surveying Tools 2026
# AI-Powered Delay & Disruption Cost Claims Analysis System # Project: Meridian Tower Extension - Quantity Surveyor Support Module from ai_modules import SOPADeadlineTracker from ai_modules import RFIClassifier from ai_modules import DailyReportWriter from ai_modules import CostImpactAnalyzer from ai_modules import TimelineComparator from ai_modules import EvidenceAggregator # Initializing delay and disruption claim preparation workflow ✓ SOPADeadlineTracker: Baseline schedule loaded (Original vs. Revised comparison) ✓ RFIClassifier: 847 RFIs analyzed - 312 critical path impacts identified ! CostImpactAnalyzer: 23 cost items require manual review - high uncertainty values ✓ DailyReportWriter: 156 site diary entries cross-referenced with delay events ✓ TimelineComparator: Float consumption analysis complete - 18 days critical delay documented ✗ EvidenceAggregator: 4 missing contractor correspondence files - claim strength reduced by 12%
At the weekly cost report meeting on a Friday afternoon, your project manager wants to know what the delay is costing — in dollars, this week. Prolongation cost calculation is methodical work: time-related preliminaries, staff costs, plant on-hire, and insurances tracked against the extended programme period. It’s also exactly the kind of structured, repetitive analysis that AI handles efficiently.
The key is giving the AI clean input data. If your preliminaries are tracked in a cost management system like Procore (from $375/month per project) or CHEOPS (pricing on request, mid-tier enterprise), export your weekly cost reports by cost code. Then use an AI tool to apply the appropriate cost rate to the extended period.
Here’s a structured prompt template for prolongation cost analysis:
Use this template:
I am preparing a prolongation cost claim under [CONTRACT TYPE e.g. AS4000-1997] for a [DURATION] week extension of time. The following are the weekly time-related preliminary cost rates extracted from our cost report for the period [DATE RANGE]. Calculate the total prolongation cost for each cost head, apply the agreed staff cost uplift of [X]%, and identify any cost heads where the weekly rate has fluctuated by more than 10% — flag these for QS review. Present results as a formatted table with cost head, weekly rate, number of weeks, total cost, and any variance flag.
PROLONGATION COST REGISTER — STRUCTURED INPUT FORMAT
Project: [PROJECT NAME] | Contract: [CONTRACT NUMBER]
EOT Granted (weeks): [NUMBER]
Delay Period: [START DATE] to [END DATE]
COST HEAD | COST CODE | WEEKLY RATE ($) | WEEKS | TOTAL ($) | FLAG
-----------------------|-----------|-----------------|-------|-----------|------
Site Manager | 1100.01 | 4,200.00 | 12 | 50,400 |
QS - Resident | 1100.02 | 3,800.00 | 12 | 45,600 |
Tower Crane - Hire | 2300.01 | 8,500.00 | 8 | 68,000 | *
Site Shed Hire | 2400.01 | 1,100.00 | 12 | 13,200 |
Public Liability Ins. | 3100.01 | 620.00 | 12 | 7,440 |
Claude 3.5 Sonnet is best suited for this structured data analysis — its ability to handle tabular inputs and apply consistent logic across cost heads makes it more reliable for this task than general web-browsing AI tools.
how to structure preliminary cost codes for claims
3. Building Compliant Claim Narratives Faster with Construction Claims Preparation AI
At 9am on a Monday, your commercial manager drops a submission deadline on your desk: the formal delay and disruption claim needs to go out in 10 working days. You have the evidence. You have the numbers. But writing a 60-page claim document that reads coherently, follows the contract’s notice provisions, and anticipates the project director’s counter-arguments takes time you don’t have.
This is where AI writing assistance adds real value for QS practitioners — not to replace your professional judgment, but to accelerate the drafting of structured claim narratives.
ChatGPT-4o or Microsoft Copilot for Word (included in Microsoft 365 Business Standard, from $15/user/month) can take your delay event schedule, your prolongation cost register, and your EOT entitlement analysis and turn them into structured draft sections following AACE International Recommended Practice formats or the SCL Delay and Disruption Protocol.
| Claim Section | Manual Drafting Time | AI-Assisted Drafting Time | Key AI Input Required |
|---|---|---|---|
| Executive Summary | 4–6 hours | 45 minutes | Delay event schedule + EOT granted |
| Entitlement Analysis | 8–12 hours | 2–3 hours | Contract clauses + notice dates |
| Prolongation Cost Schedule | 6–8 hours | 1–2 hours | Cost register + weekly rates |
| Programme Narrative | 10–15 hours | 3–4 hours | Baseline vs as-built comparison |
| Appendix Index + Cross-Refs | 2–3 hours | 20 minutes | Document register CSV |
The AI won’t get the legal nuance right without your direction — but it will produce a first draft in the correct structure that you edit rather than write from scratch. That’s the difference between a 10-day deadline and a 3-day crunch.
4. Validating As-Built Programmes and Identifying Concurrent Delay with AI Quantity Surveying Tools 2026
During a post-project dispute review, the respondent’s expert claims concurrent delay negates your client’s EOT entitlement for six weeks of the extension period. You need to disprove it — fast. Concurrent delay analysis requires comparing baseline logic, as-built sequences, and cause-and-effect chains across months of programme data. Done manually, this is a forensic exercise that takes weeks.
AI tools can accelerate the pattern recognition component. Feed ChatGPT-4o or Gemini Advanced (from $28.99/month as part of Google One AI Premium) a structured export of your Primavera P6 or Microsoft Project as-built schedule alongside your delay event log, and ask it to identify periods where both employer-risk and contractor-risk delay events overlapped.
Gemini Advanced (best suited for users already in the Google Workspace ecosystem) handles large structured data exports well and can produce a concurrent delay matrix identifying each week of the delay period, the active delay causes, and the dominant risk owner.
The AI output is a starting point for your expert analysis — not a replacement for a proper forensic programme review. But it significantly narrows the field, letting your forensic programmer focus on the contentious weeks rather than reviewing the entire extended period from scratch.
Frequently Asked Questions
Can AI actually replace a QS on a delay and disruption claim?
No — and any tool that implies otherwise is overselling. AI accelerates the document collation, drafting, and calculation components of a claim, but QS professional judgment is still required to classify risk events correctly, assess entitlement under the specific contract, and stand behind the numbers in a formal submission or adjudication. Think of it as a highly capable junior analyst, not a replacement.
What’s the best AI tool for construction delay cost claims in 2026?
There’s no single best tool — it depends on the task. Claude 3.5 Sonnet handles large document ingestion well. ChatGPT-4o is strong for structured drafting and table generation. Microsoft Copilot for Word is the most practical for QS professionals already working in Word-based claim documents. Use them in combination rather than picking one.
Is AI-generated claim content admissible in adjudication or arbitration?
The claim document you submit is yours — the professional who signs it takes responsibility for its accuracy and completeness. AI is a drafting and analysis tool, not an author. As long as your QS reviews and validates every figure and assertion before submission, the process used to draft it is no different to using a spreadsheet. Always disclose your methodology to your legal team if the matter proceeds to formal dispute resolution.
How do I ensure the AI doesn’t hallucinate dates or cost figures in my claim?
Always provide structured, clean source data — don’t rely on the AI to recall figures from memory. Upload the actual documents or data exports, and instruct the model to cite the source document and page or row number for every figure it references. Then verify a 10–15% sample of the output against the originals before incorporating it into your claim.
Conclusion: Three Things You Can Do This Week
Delay and disruption claims are document-intensive, time-critical, and commercially high-stakes. The QS professionals who recover the most from extended projects in 2026 will be the ones who build their contemporaneous record structures during the works — not after the dispute crystallises.
The three most actionable takeaways from this article:
- Start your delay event register now, even if you’re still on-site. A structured CSV updated weekly is worth more at claim time than three months of unindexed daily reports.
- Use Claude or ChatGPT-4o to ingest and cross-reference your existing records — one afternoon of structured prompting can replace days of manual reconciliation.
- Draft your prolongation cost schedule in AI-compatible formats from the beginning — clean cost code exports mean you can generate a compliant prolongation schedule in hours, not weeks.
If you want to stay ahead of how AI is reshaping commercial and contractual practice for QS professionals, the ConstructionHQ newsletter covers practical workflows, tool reviews, and claim strategy every fortnight.
subscribe to the ConstructionHQ newsletter
how to write a contractually compliant EOT claim