I have created an attended robot which would run in a client’s local machine(personal laptop) where he is able to run the robot and enter the parameters in ‘Input dialog’. Also, end to end execution takes around 2-3 hours. So client has to sit in front of machine and look at his screen without being able to perform any personal tasks in his own local machine.
My process includes the steps where UI is involved where UI based selectors clicks and also key board shortcuts are there. Login screens and multiple screens are also involved.
I was wondering if I can run such robot in the user machine whilst user will be able to perform his/her daily tasks.
Also what are the chances on performing the tasks with respect to what information I have shared above?
I figured out Simulate window properties can be toggled to achieve the results but since robot needs to run on the user machine and user also needs to work on his machine actively throughout the entire execution (2-3 hours), I find using simulate approach kinda shaky!!!
@hacky Even with Picture in Picture Property to enable users to Perform UI Actions of their own, If you are again interacting with an element which is supposed to be under the actions of the attended robot, then it might create some inconvenience in the process. Also you need to know that it’s an experimental feature added and they are still identifying different bugs which arise under different circumstances. If it’s a real project Scenario, i would Suggest not to use it yet.
This sounds like it should be running as unattended. If the process needs to be triggered by a user and the information is gathered upfront then you can gather the user input using a very simple attended bot which pushes the data into a queue for an unattended bot to process.
This way there is 0 impact on the end-user.
You could use a Custom Input form to perform the gathering of information upfront in one form:
I understand that you are currently running the process on the user local machine, but if the process take 2-3 hours then this isn’t the correct way to run the process. The correct way would be:
User kicks off Attended bot from UiPath Robot Tray
Attended bot asks the user for the required input
Attended bot puts this information in a queue and triggers an unattended bot
a) Attended bot finishes - User gets machine back. b) Unattended bot runs for 2-3 hours performing the required tasks.
Unattended bot lets the user know when it is finished (email?)
Any alternative method will still cause inconvenience to the user, even if it is minimised by new techniques (i.e. Picture-in-Picture) it will still be inconvenient. If you do not have unattended bots available then I think the question should be “How do we enable an environment to do complex processing?” and if not then “Have we chosen the correct process to automate given our current environment?”.
It all makes sense to me now. But then still, if we have process running in user machine following the steps 1,2,3,4,5, user machine will still be engaged with my process and they cannot work on their machine right?
Also what inconvenience can be causes in Picture in Picture process?
What I was wondering was, (considering in can have the UiPath studio version 2020.4.1 )+ version in user machine and using Picture in Picture setting to execute.
As soon as user is able to trigger the robot, PIP session will be progressed and user can enter their credentials as a sign of giving out permission, and then user can enter the inputs to the PIP session. and let the robot get started.
As the execution is progressed in the PIP session, user need not worry about interrupting the robot and have their own work in local session. As the execution is completed, user can recieve the mail or acknowledgement saying that the process is completed successfully.
Can you share your thoughts on this process approach?
So the attended bot will finish once they have submitted the required information and therefore the user will get the machine back as the process will then run on an unattended bot on a VM somewhere else.
I think your process for PIP would potentially work, however a key thing to remember is that it is still an experimental feature and as @supermanPunch mentioned, I would advise against putting it into production scenario.
we dont need the unattended bot to run in some VM, we need to run it in the USERS LOCAL MACCHINE.That is what I was trying to communicate.
Also, I agree to it that since it is a experimental source, its not recommended to use in production set-up.
As I mentioned I was able to look as SIMALATION toggle too, in type into, click and hotkey activities, but since bot needs to run on in the user machine(and not in some VM), I find this approach kinda shakky too!!!
This is the key player @hacky
You see we can use BACKGROUND PROCESS template to work on this especially being a attended bot
But that process doesn’t allow UI interaction
So we cannot afford that…
Either way we can try with Unattended bot , but as you say
We won’t be able to go to that as well
And the final possibilities of PIP is also not available in enterprise
Now the ball is your court
—if you want the process to be run in user machine find a way to access the application without UI interaction like using API
—we need to have a unattended bot
—your user must be ready to run that process at times apart from his working hours so that he can use the machine at his/her hours
I agree with everyone’s point of view that PiP is still in beta.
However, can you try it on a particular machine to test if it fits your processes?
I have tried several processes of my company with PiP, then I realized that it could not open 2 the same browser at the same time on the real machine and PiP
And I want to ask if you are running the robot on a schedule or if the user has to activate it on UIPath Assistant. If your user has it enabled on it then I think he can just work on PiP and other work without much effect.
Why don’t you give it a try and give me 1 comment on it
Here is a demo of PiP, I cannot capture my company process, please excuse me
That makes complete sense! Thanks for your valuable inputs!
I have the same views as what you guys have mentioned. So, as I have the following conditions to satisfy:-
Our process runs on the user local machine and also want to provision the user to work on his own parallelly.
I was able to explore PIP too, but the views provided here regarding the efficiency, looking at the current version and its experimental nature, looks difficult.
Regarding the simulate activities, its kinda tough choice to let to user mess around their machine when there are alot of dependencies on UI click and UI hotkeys and UI type into, etc
Also I need to have a attended bot and want the user to throw in the parameters as they trigger the bot from the robot TRAY. So I guess we are not in the situation to switch back to unattended robot. NO OPTION TO HAVE UNATTENDED BOT.
Cannot use the API calls as the UI based automation which is been used is the kind which cannot be handled well using APIs.
User is able to run the process at the times apart from working hours and in fact user is doing that for now. but due to obvious reasons user could find it inconvenient bcoz if the number of iterations increase, the execution might go up to 6-7 hours and that makes it even more difficult to digest such cases.
I have done my part of communicating with the team regarding all the possibilities and looks like we still need to explore more, so to conclude, can we all just have the final word on this?
I am not sure how do I proceed on this.