How do I find repeat failures in my work order history by asset, problem and vendor with AI?

Short answer

Use Copilot in Excel or Gemini in Google Sheets, with Claude or ChatGPT: Copilot or Gemini adds callback columns to your CMMS work order export that link each corrective job to an earlier one on the same asset inside a window you choose, then summarizes repeats by asset, problem type and vendor; Claude or ChatGPT turns close-out notes into consistent cause categories you review. When a contractor closes a job and the same fault is back a week later, most systems log it as a brand-new work order, so callbacks disappear into the job count and the asset that keeps failing never stands out. Export your work orders from the CMMS (or the spreadsheet you track them in), have Copilot in Excel or Gemini in Google Sheets link each corrective job to the previous one on the same asset inside a callback window you choose, and have Claude or ChatGPT sort the close-out notes into consistent cause categories. The result is a monthly list of assets, problem types and vendors with the most repeats. Root cause, vendor conversations and repair-or-replace decisions stay with your maintenance team. A person still checks: every repeat link against the work orders themselves (same asset, really the same fault), and any root cause, vendor or replace-versus-repair decision by your maintenance lead and the equipment specialists.

AI tie-in

Tool: Copilot in Excel or Gemini in Google Sheets, with Claude or ChatGPT

Copilot or Gemini adds callback columns to your CMMS work order export that link each corrective job to an earlier one on the same asset inside a window you choose, then summarizes repeats by asset, problem type and vendor; Claude or ChatGPT turns close-out notes into consistent cause categories you review.

Prompt 1

Here are the column headers and five sample rows from my CMMS work order export [paste headers and rows with names removed]. The table is called WO. Write Excel formulas (or a step-by-step Copilot plan) that add these columns: PrevWO (the most recent earlier corrective work order on the same asset ID), DaysSincePrev (days from that job's completion date to this job's request date), and Repeat (YES when DaysSincePrev is at or under the callback window in cell Settings!B1, which I set). Then build a summary by asset ID, by problem code and by vendor with counts of jobs, repeats and repeat rate. Use only columns in my data, do not invent values, and leave blanks where a date is missing.

Prompt 2

Here are the close-out notes for the work orders marked Repeat [paste work order number, asset ID, problem code and close-out note]. For each row, put the note into one of these categories, using only what the note says: adjusted or reset, part replaced, cleaned, temporary fix, waiting on parts, no fault found, cause not stated. Add a short reason quoting a few words of the note. Then list assets with three or more repeats and, for each, the categories in date order. Do not diagnose the equipment and do not say whether a vendor did poor work.

Human check: Every repeat link against the work orders themselves (same asset, really the same fault), and any root cause, vendor or replace-versus-repair decision by your maintenance lead and the equipment specialists.

The problem

Find repeat failures in your work order history by asset, problem and vendor with AI. This is a common, time-consuming task for a facilities manager.

What this workflow does

When a contractor closes a job and the same fault is back a week later, most systems log it as a brand-new work order, so callbacks disappear into the job count and the asset that keeps failing never stands out. Export your work orders from the CMMS (or the spreadsheet you track them in), have Copilot in Excel or Gemini in Google Sheets link each corrective job to the previous one on the same asset inside a callback window you choose, and have Claude or ChatGPT sort the close-out notes into consistent cause categories. The result is a monthly list of assets, problem types and vendors with the most repeats. Root cause, vendor conversations and repair-or-replace decisions stay with your maintenance team.

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

Your checkpoint: The sheet links work orders and counts repeats; it does not diagnose equipment or judge a vendor's work. A repeat flag only means the same asset came back inside your window, which may be a new, unrelated fault. Check each link against the work orders, and leave root cause, contract and repair-or-replace decisions to your maintenance lead and specialists.

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

Export corrective work orders for the last 12 months from your CMMS to Excel or CSV: work order number, asset ID, location, problem code or description, request date, completion date, vendor or technician, cost and close-out note. Remove requester names before pasting anything into an AI tool.
Decide your callback window (for example the warranty period in your service contracts) and type it into a Settings cell. The window is your policy; the AI does not pick it.
Have Copilot in Excel or Gemini in Google Sheets add PrevWO, DaysSincePrev and Repeat columns that find the most recent earlier corrective job on the same asset and mark a repeat when the gap is inside your window. Spot-check ten rows by hand before trusting the column.
Paste the repeat rows into Claude or ChatGPT and have it place each close-out note into a fixed set of categories (adjusted or reset, part replaced, temporary fix, no fault found, cause not stated). Notes that say nothing useful are the first thing to fix in your close-out rules.
Build a summary by asset, problem code and vendor: jobs, repeats, repeat rate and total cost. Sort by repeats and take the top of the list to the monthly maintenance or vendor meeting with the work orders in hand.
Going forward, reopen or link the original work order when a fault returns inside the window instead of opening a fresh one, and require a cause note before a job can be closed. Rerun the export each month and compare.

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

Here are the column headers and five sample rows from my CMMS work order export [paste headers and rows with names removed]. The table is called WO. Write Excel formulas (or a step-by-step Copilot plan) that add these columns: PrevWO (the most recent earlier corrective work order on the same asset ID), DaysSincePrev (days from that job's completion date to this job's request date), and Repeat (YES when DaysSincePrev is at or under the callback window in cell Settings!B1, which I set). Then build a summary by asset ID, by problem code and by vendor with counts of jobs, repeats and repeat rate. Use only columns in my data, do not invent values, and leave blanks where a date is missing.

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 sheet links work orders and counts repeats; it does not diagnose equipment or judge a vendor's work. A repeat flag only means the same asset came back inside your window, which may be a new, unrelated fault. Check each link against the work orders, and leave root cause, contract and repair-or-replace decisions to your maintenance lead and specialists.
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?

Copilot in Excel or Gemini in Google Sheets, with Claude or ChatGPT: Copilot or Gemini adds callback columns to your CMMS work order export that link each corrective job to an earlier one on the same asset inside a window you choose, then summarizes repeats by asset, problem type and vendor; Claude or ChatGPT turns close-out notes into consistent cause categories you review. A person still checks: Every repeat link against the work orders themselves (same asset, really the same fault), and any root cause, vendor or replace-versus-repair decision by your maintenance lead and the equipment specialists.