This is a classic data architecture problem dressed up as a pivot table frustration, and the root cause is almost certainly not duplicates in your source data but a misunderstanding of how Excel's data model handles relationships. You're not alone in hitting this wall, and the fix is simpler than the error message suggests.
The core issue is that you're trying to use a single slicer across multiple unrelated pivot tables. Excel's traditional pivot table and slicer system requires a shared data model or at least a common lookup table to connect those separate datasets. When you're pulling sales data directly from your CRM, each dataset exists in its own silo, different time periods, same salesperson names. The slicer cannot magically bridge those silos unless you tell Excel how they relate. The "duplicates" error you're seeing is Excel telling you that it found multiple instances of the same salesperson name across your tables and can't decide which one to use as the filter anchor. It's not a data quality issue; it's a relationship design issue.
Here's the practical fix: Instead of trying to filter four separate pivot tables with one slicer, consolidate all your datasets into a single table with a "Period" column. Append your datasets vertically, each row gets a salesperson name, a sales amount, and a column that identifies which time period it belongs to. Then build one pivot table from that combined table. Your slicer for salesperson will now filter everything because it's all in one logical dataset. You can then use the Period column as a filter or a secondary slicer to toggle between time frames. This approach also eliminates the relationship headache entirely because you're not asking Excel to guess how two separate tables connect.
If appending isn't feasible because the datasets are too large or come from different sources you can't merge, you need to create a shared lookup table of salesperson names, a small, deduplicated list, and connect each dataset to that lookup table using Power Pivot or the Excel Data Model. That gives you a proper star schema where one slicer on the lookup table filters all related fact tables simultaneously. This is more advanced but necessary when datasets must remain separate. Either way, stop fighting the slicer error and start designing your data structure to match how you actually want to filter. The tool isn't broken; your approach just needs to align with how relational filtering works.