Yeah that's definitely the way to go if there's already a helm chart. If it's just raw manifests I'd rather convert to HCL than build a helm chart to feed into TF, but that's just because I know TF better, most of my configuration is already there, so I'd prefer to keep abstraction boundaries small. But it's nice to have the options.
It’s fairly simple to set up a basic helm chart around existent yaml specs, then progressive parameterization of that yields a nice managble object.