# AI Project Closeout > AI project closeout uses document extraction and agents to collect, classify, and verify the O&M manuals, warranties, as-builts, and lien waivers required at turnover. It tracks outstanding items by subcontractor. - Stages: Closeout - Concepts: Document Extraction, Agents - Published: 2026-08-28 - Canonical: https://buildr.com/library/ai-project-closeout ## Why it matters in construction Closeout is where the last 5 to 10 percent of contract value sits, and it is the phase nobody wants to work on. Substantial completion hits, the super and PM move to the next job, and a PE inherits a spreadsheet of 400 required documents pulled from Division 01 and every technical spec section. Retainage does not release until the owner accepts the package, and the owner will not accept the package until the mechanical sub sends the correct warranty letter for the third time. Closeout can run for months after substantial completion. That leaves retainage on the balance sheet and keeps a PE tied to a job that is no longer billing. ## How it works 1. **Build the requirements list.** A model reads Division 01's closeout section plus the closeout submittal paragraph in every technical section and produces the list of required documents by trade and sub. 2. **Classify incoming documents.** As subs send PDFs, the tool tags each one as O&M, warranty, as-built, certification, or lien waiver and matches it to the requirement and sub it satisfies. 3. **Verify.** Extraction checks the content, not the file name. A warranty letter needs the right project name, a start date at substantial completion rather than install, a duration that matches the spec, and a signature. A lien waiver needs to match the pay app amount. 4. **Track and chase.** A dashboard shows percent complete by sub. An agent sends each sub a specific list of what is outstanding and what got rejected and why. 5. **Assemble.** The final package goes together by spec section with an index, ready for the owner and the facilities team. Every "complete" checkmark should link to the document and the spec paragraph it satisfies. ## Example in practice Consider a commercial GC reaching substantial completion on a $28M office building with 34 subs and $1.6M in retainage outstanding. The spec generates 380 closeout requirements. Within a week, subs have sent 210 documents, which the system classifies. It rejects 14: six warranties dated from installation rather than substantial completion, five O&M manuals that are product cut sheets rather than manuals, and three unsigned lien waivers. Each sub receives a list explaining the issue. The electrical sub, 40 percent complete on 52 items, receives a weekly reminder and a note to the PM at day 21. In this example, the owner accepts the package at week nine rather than month six, releasing $1.6M in retainage a quarter earlier. ## Go deeper - [Successful Project Closeout: The Comprehensive Guide](/blog/project-closeout.md) - [How to Connect Your Preconstruction Workflow to Procore Without Manual Data Entry](/blog/connect-preconstruction-workflow-to-procore.md) ## How Buildr applies this Buildr organizes closeout documents and warranty information by project and tracks outstanding items from each subcontractor through turnover. See [Buildr Handover](/handover). ## Related terms - [Structured Data Extraction](/library/structured-data-extraction.md): Structured data extraction uses AI to turn unstructured construction documents, such as sub proposals, RFPs, and specs, into typed fields like line items, prices, exclusions, and dates that can be sorted, compared, and loaded into an estimate or CRM. - [AI Agent](/library/ai-agent.md): An AI agent uses a language model and connected tools to carry out multi-step tasks. For example, it can read a bid invitation, check it against go/no-go criteria, and create an opportunity in the CRM. - [Agentic Workflow](/library/agentic-workflow.md): An agentic workflow is a business process in which AI agents handle defined steps, such as intake, extraction, and record updates, while people review the work at checkpoints. A preconstruction team might use one for bid intake or CRM upkeep. - [OCR vs. LLM Document Understanding](/library/ocr-vs-llm-document-understanding.md): OCR converts scanned construction documents into text. LLM document understanding reads that text, or the page image itself, and interprets its meaning. Precon tools often use both: OCR handles the pixels and the language model interprets the content. - [Human-in-the-Loop](/library/human-in-the-loop.md): Human-in-the-loop is an AI design pattern in which a person reviews, corrects, or approves the model's output before it takes effect. In construction, an estimator still owns the award and a BD lead still approves a pipeline update, while the AI handles reading and drafting. ## FAQ ### What closeout documents can AI actually read? O&M manuals, warranty letters, as-built drawings with revision clouds, certificates and inspection reports, lien waivers, and training records. Anything that arrives as a PDF or email attachment can be classified and checked against the spec requirements. ### Can it chase subs for missing documents? Yes, within limits. An agent can send reminders listing exactly what is outstanding per sub. A person should still make the call when a sub is unresponsive or when retainage release is on the line. ### Why does closeout take so long in the first place? It often happens after the team has moved to the next job, subcontractors are focused elsewhere, and the requirements are scattered across many spec sections. Without a clear owner and list, the process can drag.