> ## Documentation Index
> Fetch the complete documentation index at: https://docs.contractor-co-pilot.com/llms.txt
> Use this file to discover all available pages before exploring further.
# Create a project from uploaded construction documents
> Upload specs, blueprints, or subcontractor quotes and let AI extract the scope — then review, refine, and launch your project in one place.
Instead of manually reading through a bid package and re-entering project details, you can upload your documents directly into Contractor Co-Pilot and let AI do the heavy lifting. This workflow is particularly useful when a client sends you a set of plans, a spec book, or a subcontractor quote that you need to turn into a project quickly.
**Supported document types:** PDF, DOCX, DOC, XLSX, and XLS files up to 50 MB each. For best results, use text-based PDFs rather than scanned images. Scanned documents are supported but may produce less accurate extractions depending on scan quality. Architectural plans in PDF format work well; CAD files (.dwg, .dxf) are not currently supported.
1. Go to **Projects** in the left sidebar
2. Click **New from Documents**
The page opens with a document upload zone on the left and a project details form on the right. You'll fill in both sides as you work through this workflow.
Drag and drop your files into the upload zone, or click **Browse** to select them from your computer. You can upload multiple files at once.
Each document displays its current status as it progresses:
| Status | What's happening |
| ------------- | ---------------------------------------- |
| **Pending** | Queued, waiting to upload |
| **Uploading** | File transfer in progress |
| **Uploaded** | File received, ready for analysis |
| **Analyzing** | AI is reading and extracting information |
| **Analyzed** | Extraction complete |
Once all files are in the queue, click **Upload All** to begin the transfer.
Upload everything you have — the full spec book, all plan sheets, any addenda, and the subcontractor quote if you have one. The AI builds a more accurate picture when it can cross-reference multiple documents.
Once your documents have uploaded, click **Analyze with AI** to start extraction. The AI will:
* Identify each document's type (plans, specifications, contract, quote, etc.)
* Extract the project name, scope description, and key requirements
* Pull out notes and conditions relevant to the project
* Assign a category badge to each document so you can see what it found
Analysis typically takes 30–90 seconds per document depending on file size and complexity. You can watch the status update in real time.
The AI is looking for structured information — project names, scope language, quantities, dates, and notes. It performs best on documents that use standard construction terminology and formatting.
When analysis is complete, the project form on the right side of the screen populates with AI-suggested values:
* **Project name** — Suggested based on the document title, cover sheet, or header
* **Description** — A summary of the project pulled from scope sections
* **Scope** — Detailed scope items extracted from specifications or work descriptions
* **Notes** — Additional requirements, conditions, and clarifications found in the documents
Read through each field carefully. The AI gets most projects to 80% accuracy — your job is to catch the remaining 20%.
Common things to verify:
* Is the project name specific enough? (Add the address or client name if needed.)
* Does the scope match what you actually intend to bid or build?
* Are there exclusions or alternates in the spec that should be noted?
* Did the AI miss any critical requirements buried deep in the documents?
Click on any document in the list to review the original file alongside the extracted values. This makes it easy to verify where specific information came from.
The document analysis populates scope and description fields, but some details need to come from you:
* **Client** — Select the client from your contacts, or add a new one
* **Start date and expected completion** — Add your estimated schedule
* **Jobsite address** — Enter the physical work location if it wasn't in the documents
* **Division** — Assign the project to the correct division if your organization uses them
If the client or jobsite address appeared in your documents and wasn't extracted automatically, you can copy it from the original file (visible in the document preview pane) and paste it into the form.
Before creating the project, you can convert the extracted scope into a formal estimate:
1. Click **Create Estimate from Scope**
2. The extracted scope items populate as estimate line items
3. Add quantities, unit costs, and apply your cost profile
4. Review the total and save the estimate
This gives you a complete budget tied to the project from day one, rather than adding costs manually after activation.
Hank, your AI co-pilot, can help fill in unit costs if you're not sure on pricing for specific line items. Open Hank from the sidebar and describe the work.
Once the form is complete and you're satisfied with the extracted and reviewed details:
1. Click **Create Project**
2. All uploaded documents are automatically attached to the new project
3. The project opens in **Planning** status, ready for you to assign your team and set the timeline
From here, the workflow is the same as any other project. See [Convert an approved proposal into an active project](/workflows/proposal-to-project) for guidance on activating the project and inviting your field team.
## What's next
Complete the project setup — timeline, team, budget, and first invoice.
Learn how the AI extraction engine works and how to get the best results from your documents.