First of all welcome to SQAForums. I have put some threads below which I <font color="red"> strongly </font> recommend that you read and then follow.
Do NOT ask the same question multiple times. Find the right or most appropriate forum, ask once and then wait patiently. This forum is staffed by volunteers around the world, it can take some time before some posts get answered.
No one can really give you a decent way to estimate this workload, because we have no idea how dense, complex, and correct the manual test cases are. This type of estimate is futher complicated and intertwined with your QTP skill level versus others on this board who may offer up some ballpark value.
If the manual test cases are relatively consistent in density and complexity then randomly select half dozen and automate them, keeping copious track of all time to:
3. Ask the testcase author for clarifications and corrections.
If, as is more likely the case they are not consistent in density and complexity, then first:
1. divide them into complexity/densitiy buckets (perhaps 6 at most).
2. Then randomly select one from each bucket
3. And then perform the above automation an analysis.
After automating the selected half dozen, now average the time per test case and then divide by 134. This will be a very, very rough estimate of hours per test case.
Automating a random half dozen is the only way you are going to be able to defend even this rough guestimate (as opposed to telling a manager that you got an estimate from an sqaForums.com board member).
-Hope this squishy suggestion helps, Terry Horwath