We work as a center of excellence and we develop automations centrally for other departments. All our deployments are unattended bots that are deployed and managed by us, centrally. They are all running on a virtual machine. Nothing is running on client computers or laptops. We do not have any attended use cases. The process owners are interacting with the bots.
We use the bots for a lot of reporting, including monthly and weekly reports. We are a construction company and we have a lot of reports for all kinds of things, such as construction projects, different construction sites, and various subsidiaries for regional departments. A region like Bavaria, for example, needs its reports. And there are plenty of controlling departments in all of the subsidiaries.
We also have some ticketing use cases. One of them is for IT services, meaning internal ticketing. That bot regularly checks our ticket software and automatically processes some of the tickets. For example, when an employee needs rights to a specific system, the bot checks whether they fulfill the requirements and approves or declines the ticket.
Another type of ticketing use case is more about processing customer tickets. As a construction company, we also do facility management, and that means there are a lot of external customers with their own systems in which they record tickets. The tickets are not visible in our local systems so someone has to go to the external systems, export the tickets, filter them, and then tell everyone what they're supposed to do to their buildings as a result. The tickets might be about small repair jobs, for example. We run this daily and, in the morning, everyone receives an email with all the tickets that have to be done within one day, three days, one week, et cetera.
Both of the ticketing use cases are connected with SLAs. If you miss a certain time frame before processing a ticket for external customers, you have to pay a penalty fee. For the internal tickets there are SLAs for internal tracking purposes. Because those tickets have to be processed within two hours, that bot runs every two hours and checks for new tickets.
Another IT services use case is for getting access rights to local drives.
We also have many recurring processes. For example, in HR they have to go to the system and confirm a process. It’s a necessary evil which is probably due to the legacy systems we have. Someone defined this process a long time ago and it still has to be done.
We also have use cases in finance and treasury. They are not tickets, but they process requests from employees. For example, they can request cash on a specific card and the bot will check the emails and then basically transfer data from an email, or from a PDF form attached to an email, and enter it into the finance system.
One last type of use case is where the bot works as an interface between systems. Data has to be exported from one system and imported to another system and there is no existing API. The bot exports and imports the data. We have one bot that exports PDF documents and sends them to an email interface. It defines a specific subject and then attaches the file. That file will automatically be uploaded to another system. Or the bot may log in to a system and upload the document. These use cases are due to the fact that there is no interface between two systems and they're either not big enough to develop an API or they may involve an external customer system and the customer has no interest in providing an API.