Custom GPT: Company Scope Template Library for Estimators

What This Builds

You'll build a Custom GPT trained on your company's scope letter templates, standard exclusion language, and bid invitation formats. Any estimator on your team can open it and get a properly scoped, company-voice letter for any trade in 2 minutes, instead of hunting through old folders for the last version of the template or starting from scratch.

The result: consistent scope language across your team, institutional knowledge that doesn't leave when a senior estimator does, and 20-40 minutes saved every time someone writes a scope letter.

Prerequisites

• {{tool:ChatGPT.plan}} subscription ({{tool:ChatGPT.price}}; the Custom GPT builder requires Plus or higher)
• 5-10 examples of your best scope letters from past bids (any format: Word, PDF, email)
• Your standard exclusion language (if you have a list)
• Time: 1-2 hours to build; 2 minutes per use afterward

The Concept

A Custom GPT is like training a new employee who already knows your company's style. You give it examples of how you write, what you typically include or exclude, and how you like things formatted. After that, anyone on your team can ask it "write a mechanical scope letter for a 3-story medical office" and get something that sounds like your company wrote it, not a generic template.

Unlike a Word template that you copy-paste and fill in, this GPT understands your scope decisions and can apply them to new project types it hasn't seen before.

Build It Step by Step

Part 1: Gather and organize your source material

Before opening the GPT builder, collect:

1. 5-10 scope letter examples: Pull your best scope letters from the last 2 years across different trades and building types. Good examples = better output. Aim for: mechanical, electrical, drywall, concrete/structural, and at least 2 other trades.

2. Your standard exclusion language: If your company has a boilerplate list of things you always exclude from the GC scope (e.g., "Owner-furnished equipment hookups not included unless noted," "Sales tax is excluded"), copy this into a text file.

3. Your formatting preferences: Note whether you use numbered lists or paragraphs, whether you include a signature block, and whether you use specific headers like "Scope of Work" vs. "Included in Bid".

4. Common project types: List the 5-7 building types you bid most often (e.g., office TI, healthcare, warehouse, multifamily, education).

Part 2: Create the Custom GPT

1. Log into {{tool:ChatGPT.url}} with your Plus account
2. Click Explore GPTs in the left sidebar, then Create at the top right
3. You'll enter the GPT Builder with a "Create" tab and a "Configure" tab. Use Configure for more control
4. Set the Name: "Estimating Scope Assistant: [Your Company Name]"
5. Set the Description: "Drafts bid invitation letters, scope letters, and exclusion language in [Company Name]'s style for any trade and project type."

Part 3: Write the system instructions

In the Instructions field (the most important part), paste this template and customize the bracketed sections:

You are an estimating scope assistant for [Company Name], a commercial general contractor based in [city/region] specializing in [building types, e.g., office, healthcare, warehouse].

Your job is to draft subcontractor scope letters, bid invitation letters, and scope inclusion/exclusion lists in [Company Name]'s voice and format.

COMPANY STYLE:
- [Describe your tone: e.g., "formal and direct" or "professional but not overly legalistic"]
- [Describe your format: e.g., "use numbered lists for scope items, paragraph format for introduction and closing"]
- [Any specific language: e.g., "always reference the project address, not just the name"]

STANDARD EXCLUSIONS (always include these unless instructed otherwise):
- [paste your standard exclusion list here]

STANDARD INCLUSIONS FOR BID INVITATIONS:
- Bid bond requirement (if applicable): [your policy]
- Insurance requirements: [your minimums]
- Submission format: [your requirements]

WHEN ASKED TO DRAFT A SCOPE LETTER:
1. Ask for: trade, project name, project type, project size (SF), location, and bid date
2. Produce a complete scope letter with: introduction, scope inclusions, scope exclusions, special requirements, submission requirements, and contact placeholder
3. Match the format of the uploaded example letters

Do not invent scope items that weren't implied by the project description. If you're uncertain about a scope boundary, flag it with [CONFIRM WITH PM] rather than guessing.

Part 4: Upload your example files

1. In the GPT Configure page, scroll to Knowledge
2. Click Upload files
3. Upload your 5-10 scope letter examples. PDF or Word both work
4. Upload your standard exclusion language document
5. Click Save after uploading

Part 5: Test with realistic requests

Before sharing with your team, test it thoroughly:

1. Ask for a scope letter for a trade you uploaded examples for: "Draft a mechanical scope letter for a 60,000 SF medical office in Nashville. Bid due April 15."
2. Ask for a trade you didn't upload examples for: "Draft a curtain wall scope letter for a 12-story office tower."
3. Ask it to adjust: "Make the exclusions section shorter" or "Add a paragraph about our preferred sub's scope of supply vs. install"

For each test, check:

• Does the format match your company's style?
• Are your standard exclusions included?
• Is the language appropriate for the trade and project type?
• Did it flag any uncertain scope items correctly?

Refine your system instructions based on what you see.

Real Example: Full Workflow

Setup: You've uploaded 8 scope letters from past bids (2 mechanical, 2 electrical, 2 drywall, 1 structural steel, 1 roofing). Your standard exclusion language is in the knowledge base.

Input (what an estimator types): "Draft a plumbing scope letter for a 4-story, 90,000 SF office building in Atlanta. Owner is [Owner Name]. Bid due May 3. The plumbing scope includes all domestic water, sanitary, and storm within the building. No site utilities."

Output (what the GPT produces): A complete scope letter in your company's format: introduction with project name and bid date, numbered scope inclusions (domestic water, sanitary, storm, within-building only), scope exclusions (site utilities, medical gas, compressed air, Owner-furnished equipment), submission requirements, and contact placeholders.

Time saved: 25-35 minutes vs. finding and editing an old template.

What makes it better than a template: The GPT applied the "within-building only" constraint you mentioned to set the scope boundary correctly. A plain template wouldn't know to do that.

What to Do When It Breaks

• Scope letter doesn't match your company style → Your instruction examples may be overriding the uploaded files. Add more specific formatting instructions to the system prompt (e.g., "Use the exact header format from the uploaded mechanical scope letter from 2024").
• It invents scope items you didn't include → Add to instructions: "Do not assume scope items beyond what the project description clearly implies. When in doubt, exclude and flag."
• It keeps including wrong exclusions → Update your standard exclusion list in the knowledge base and re-upload. The knowledge base overrides general ChatGPT knowledge.
• Output is too long / too short → Add a length guideline to instructions: "Scope letters should be 1-2 pages. Keep scope lists to 8-15 items maximum."

Variations

• Simpler version: Instead of a Custom GPT, create a single saved ChatGPT conversation with your examples pasted in and share the conversation link with your team. Less elegant but zero setup.
• Extended version: Add a second Custom GPT for bid analysis and reporting, using the same approach with bid review templates. Connect them both under a shared team workspace (ChatGPT Teams plan).

What to Do Next

• This week: Build and test the GPT with 3-5 real scope letter requests; refine the instructions based on what comes back
• This month: Share with your estimating team; collect feedback on what language isn't matching your voice and update the knowledge base
• Advanced: Add your company's standard subcontract terms and conditions to the knowledge base so the GPT can flag when a requested scope item has contract implications

Advanced guide for Construction Cost Estimator professionals. These techniques use more sophisticated AI features that may require paid subscriptions.