CSV import handles non-ASCII names without mangling them

What we fixed

CSV uploads were being parsed as Latin-1 instead of UTF-8 on the import path, so names with accents and any non-Latin characters arrived mangled. Exports were already clean, but imports weren't.

What changed

• The import parser now honors the BOM and falls back to UTF-8.
• Imports of CSVs without a BOM still get UTF-8 by default (was Latin-1).
• The pre-import preview shows the first 10 rows so you can spot encoding issues before committing.

If you were affected

Re-importing the same file should now produce the correct names. If you want help cleaning up records that landed mangled, support can run a one-shot fixup for your org.