Hi Neeraj,
Re Q1. Process Mining isn't mandatory; however it is usually the place to start as it will enable you to have a data driven approach to determine that you've found the best Task to focus on mining.
Think of it that Task Mining is the detail e.g. screen, field interaction level, whereas Process Mining is essentially lots of Tasks strung together. Process mining will help you find the bottlenecks or inefficient tasks then you drill down.
Re Q2. Like most things start small. Record 1 person perform the task once. Use that to map the ideal "happy path", once comfortable that your approach is working well, then increase the volume. So record that 1 person for say 1hr (assuming they've performed the task 10-20 times). Analyse the data – you'll see the different paths they've taken. Then continue increasing both the duration and number of people if applicable.
If you jump into the deep end and record lots of staff for many hrs/days you will have a mass of data that's really difficult to map/analyse.
Going back to your original question the more hrs/days/people (i.e data) you record the more comprehensive the data will be. However, this also makes the task of analysing the data larger and more complex.
I would also say start with a simple(ish) Task if possible. Again, if you're new to Task Mining and pick a complex Task (which let's face it, this is typically what needs to be analysed) it can be a massive challenge.
Hope this helps
------------------------------
Andrew Beard
Senior Consultant
Blue Prism
Australia/Melbourne
------------------------------
