I'm not sure exactly what you're asking, but here's some additional advice.

If you don't need XML files now, skip that step. (You might need them if you have XSLT or CSS files to apply to the XML to create HTML, or if you need an archive for the future.) If that's not currently the case, skip the intermediate step and just generate HTML.

Other than that, the data format for your input files and inside your program will dictate what kind of XML creator you can use. If it's possible that you can use the same intermediate format in steps 1 and 3, do that. That means you'll have a function to interpret data format #1 and another to interpret #3. You'll probably have one to format XML (or just HTML).

Getting the intermediate data format right is the trickiest part. Good luck.