A hybrid bot solution is one in which a transaction is split up to run on two different desktops: one part of the transaction executed by an attended bot running on the user’s local desktop, and the second part executed by an unattended bot running on a dedicated VM’s desktop. The typical use case starts when a user, after navigating to some relevant section of a line-of-business application, clicks our Ratchet-X magic button which instructs a local attended bot to execute an action. This action gathers data from relevant fields and submits it as a job to Ratchet-X Cloud RPA, where a related action is then executed by an unattended bot running on a remote VM. Once the local bot submits the job to Cloud RPA, the user is free to use their desktop as usual. In other words, once the job is submitted, the user’s desktop is freed up for other tasks. As an unattended bot becomes available to process the submitted job, that bot accepts it and performs the second part of the transaction. When the second action completes, the unattended bot informs Cloud RPA, which in turn, informs the end user via a notification message that the transaction has been completed.
Let’s take a look at a couple of real world examples.
Example 1: A banking company partners with home inspector companies to help process mortgage applications. Prior to being automated, bank employees were required to navigate to an account record in their line-of-business system and, using it for guidance, manually enter the home inspection request into the partner’s website. This process involved rekeying 20 data elements, took about 15 minutes to complete, and experienced an error rate of 12%.
After implementing a hybrid bot solution, now bank employees navigate to an account record in their line-of-business system and click the magic button. When clicked, an attended bot runs a local action that reads the 20 fields from the line-of-business application and submits a request to Cloud RPA to have the inspection request entered into the partner site by an unattended bot. This process usually takes about 30 seconds. After the local action submits the job, the user is then free to use their desktop as usual. At any time, the user can check on the status of this request by clicking on an icon in the Windows system tray. The hybrid bot solution reduced the user’s direct involvement in the transaction to just 30 seconds, and the inspection request error rate dropped to 0.
Example 2: Expanding example 1 above, the bank wanted to add a new business rule to the partner website data entry process. The partner website occasionally charges a surcharge for home inspections performed on certain properties or during certain periods. When this occurs, the bank wants to pause the data entry (on the unattended bot), and prompt either the requesting user, or any other available authorized user, to approve the surcharge.
In this case, the unattended action was modified to pause and send a “Prompter” back to the requesting desktop (or another authorized user if the original requester is unavailable). The prompter “pops up” on the user’s desktop, notifies the user about the surcharge and asks if they want to accept or reject the request. If approved, the unattended bot continues and completes the automation. If rejected, the unattended bot fails the job and sends it to a rejection queue for offline handling. If no user is available to handle the prompt, the attended bot fails the job so it can be re-run later.
As you can see, hybrid bot automations can be a powerful strategy for executing longer-running transactions. Not only does it put the user in total control, it also frees up the local desktop much earlier in the process than if the entire transaction had to run locally. It also allows prompts to be routed to either the requesting user or any other user authorized to handle questions regarding a transaction, thus efficiently distributing the workload.