Get an AI first-pass discrepancy list on MEP equipment submittals, then review every line

Short answer

Use Claude or ChatGPT: The AI compares the design values and spec items you choose against the submitted product data and returns a MATCH / DIFFERENT / NOT SHOWN table with page references, then checks a resubmittal against your earlier comments. Large equipment submittals such as air handlers, chillers, pumps and switchboards take the longest to review. Upload the product data with your equipment schedule row and the spec items you care about to Claude or ChatGPT, and ask for a table that compares each design value to what was submitted, with page references and a MATCH, DIFFERENT or NOT SHOWN status. You review from that checklist; the comments and the review action stay with you. A person still checks: every DIFFERENT and NOT SHOWN row at the cited page, a sample of MATCH rows, and the review action itself.

AI tie-in

Tool: Claude or ChatGPT

The AI compares the design values and spec items you choose against the submitted product data and returns a MATCH / DIFFERENT / NOT SHOWN table with page references, then checks a resubmittal against your earlier comments.

Prompt 1

Compare this equipment submittal to our design. Inputs: (1) our equipment schedule row: [paste]; (2) the spec items I want checked, with section numbers, pasted from our project manual: [paste]; (3) the submittal product data: [upload PDF]. Output a table: item, design value and its source, submitted value and submittal page, status (MATCH, DIFFERENT or NOT SHOWN). Show both units if you convert. Do not judge whether the equipment is acceptable or code compliant, do not add requirements from memory, and write NOT SHOWN instead of guessing.

Prompt 2

Here are our review comments on the previous submission: [paste]. Here is the resubmittal: [upload PDF]. For each comment, give the page where it appears to be addressed, or write NOT FOUND. Do not decide whether a response is acceptable.

Human check: Every DIFFERENT and NOT SHOWN row at the cited page, a sample of MATCH rows, and the review action itself.

The problem

Get an AI first-pass discrepancy list on MEP equipment submittals, then review every line. This is a common, time-consuming task for a mep engineer.

What this workflow does

Large equipment submittals such as air handlers, chillers, pumps and switchboards take the longest to review. Upload the product data with your equipment schedule row and the spec items you care about to Claude or ChatGPT, and ask for a table that compares each design value to what was submitted, with page references and a MATCH, DIFFERENT or NOT SHOWN status. You review from that checklist; the comments and the review action stay with you.

How it works: AI drafts or analyzes. A person operates the tools, checks the work, and takes action.

Your checkpoint: The engineer reviews every row and owns the review action. Models misread tables, units and multi-option data sheets, and they cannot judge design intent, coordination or code compliance. Keep project documents in tools your firm has approved.

Before you start

Access to the named tool; confirm license, version and cost with the vendor.
A small, non-confidential sample to try the workflow on first.
A qualified person to review the output before it is used.

Tools and access

Follow the steps

Gather three inputs: the submittal product data PDF, the matching row of your equipment schedule, and the spec section from your licensed project manual.
Pick the items that matter for this piece of equipment, roughly 10 to 30: scheduled performance values, electrical data, dimensions and weight, connections, and the spec items you always check. A focused list beats asking the model to test every sentence of the spec.
Ask for a comparison table: item, design value and its source (schedule or spec section number), submitted value and submittal page, and a status of MATCH, DIFFERENT or NOT SHOWN. No approval verdict.
Open every DIFFERENT and NOT SHOWN row at the cited page, then spot-check a few MATCH rows, because models can read the wrong column, unit or option on a multi-model data sheet.
Decide and write the review comments yourself; the model can tidy the wording of comments you have already decided on.
On a resubmittal, paste your previous comments and ask where each one appears to be addressed, so you can check those pages first.

Try a small example

Replace the bracketed fields with your own project details. No tested sample run is published for this workflow yet.

Sample input

Fill in the bracketed [fields] in the text below with your project data.

Starter prompt

Compare this equipment submittal to our design. Inputs: (1) our equipment schedule row: [paste]; (2) the spec items I want checked, with section numbers, pasted from our project manual: [paste]; (3) the submittal product data: [upload PDF]. Output a table: item, design value and its source, submitted value and submittal page, status (MATCH, DIFFERENT or NOT SHOWN). Show both units if you convert. Do not judge whether the equipment is acceptable or code compliant, do not add requirements from memory, and write NOT SHOWN instead of guessing.

Expected output

Not published yet. Treat any output as a draft and check it against the source documents.

Check the result

Every value, citation or quantity traces back to a source document.
Missing information is marked as missing, not filled in.

Limits and recovery

The engineer reviews every row and owns the review action. Models misread tables, units and multi-option data sheets, and they cannot judge design intent, coordination or code compliance. Keep project documents in tools your firm has approved.
Check your organization’s data policy before sharing project files with third-party tools.

Stop or switch when: The source cannot be verified, permissions are missing, or the result contains unsupported facts.

Frequently asked questions

How does AI help with this?

Claude or ChatGPT: The AI compares the design values and spec items you choose against the submitted product data and returns a MATCH / DIFFERENT / NOT SHOWN table with page references, then checks a resubmittal against your earlier comments. A person still checks: Every DIFFERENT and NOT SHOWN row at the cited page, a sample of MATCH rows, and the review action itself.

How do I get an AI first-pass discrepancy list on MEP equipment submittals, then review every line?

Gather three inputs: the submittal product data PDF, the matching row of your equipment schedule, and the spec section from your licensed project manual. Pick the items that matter for this piece of equipment, roughly 10 to 30: scheduled performance values, electrical data, dimensions and weight, connections, and the spec items you always check. A focused list beats asking the model to test every sentence of the spec. Ask for a comparison table: item, design value and its source (schedule or spec section number), submitted value and submittal page, and a status of MATCH, DIFFERENT or NOT SHOWN. No approval verdict. The full step list is above.

Does this use an AI agent or regular software?

It is labeled AI assists: AI drafts or analyzes. A person operates the tools, checks the work, and takes action.

What tools do I need?

Claude, ChatGPT, Microsoft Excel. Confirm license, version and cost with the vendor, and try it on a small non-confidential sample first.

Has this workflow been tested?

Workflow guide · not independently tested. Written from vendor documentation and practitioner use. Labels apply to the steps described here, not every feature of a product. Links were checked when the page was written.

Evidence and maintenance

Written from vendor documentation and practitioner use. Labels apply to the steps described here, not every feature of a product. Links were checked when the page was written.