We see this problem every day, and it's exactly the kind of mess that makes people resent their own dashboards. A coworker changes abbreviations on a whim, typos creep in, and everything lands in one column because the report was designed for paper, not for analysis. Then someone like you, the person who actually needs the data to work, has to clean it up by hand. That is not a spreadsheet problem. That is a trust problem. You cannot build a reliable view of who is working when if the source data keeps shifting under you.
The practical reality is that traditional normalization tools, including Power Query, expect consistency. They expect "M-F" to always be "M-F" and "7am" to never show up as "7:00am" or "7;00am." When your coworker enters "12pm-8;30pm Mon & Fri; 4pm-12:30am Tues.-Thurs.," you are not dealing with a formatting issue anymore. You are dealing with a logic puzzle that changes every time she types. And you are doing it in Excel 2016 with no VBA, no Python, and no automation. That is a constraint most advice threads ignore. The standard answer, "just use Power Query M", assumes you have full control over the environment. You do not.
What you actually need is a system that adapts to the mess, not one that demands the mess be fixed first. AI-native spreadsheets are built for exactly this scenario. Instead of writing a dozen nested IF statements or regex patterns that break the moment someone writes "Tues.-Thurs." with a period instead of a comma, an adaptive model can learn the patterns in your coworker's entries. It can recognize that "M&Tu" means Monday and Tuesday, that "7am-3:30pm m-f" is the same schedule as "7:00am-11:00am M, W, F 2:00pm-9:00pm T,Th" just expressed differently, and that a semicolon in "8;30pm" is a typo, not a different time. This is not magic. It is pattern recognition applied to the one thing that makes your job hard: human inconsistency.
The real insight here is that your dashboard is not the problem. You have a clear goal, show who is working, when, and for which company, with the ability to query by time window and day. That is a solid specification. The bottleneck is the ingestion layer. If you can normalize messy schedule data in a way that tolerates variation, you free yourself from policing your coworker's typing habits. That is the transformation worth exploring. Not a fancier dashboard, not more manual cleaning steps. A smarter way to handle the data as it actually arrives.