CA offers minimal public information pertaining to the performance drain the usage of some objects and operators introduce to processing. As an example, swim lanes within a process provide an excellent means of organizing operators within a process, but they can introduce substantial performance issues. As another example, it’s better to perform verbose JavaScript execution within a Run JavaScript operator instead of within another operator’s pre- or post-execution script. As yet another example, it’s better to hard-code variables within the process dataset as opposed to creating the variables at run-time. The biggest issue for me is its lack of support for current JavaScript methods and functions, which makes scripts unnecessarily longer than they need to be. It seemed I could only rely on the methods and functions available in ECMA 1 (which was released in 1997), but that wasn’t a deal-breaker and the product is capable of extending its capabilities through the inclusion of other code libraries.

I'd like to see additional form functionality and having Python as an available common script language.

It would be good if we could have an interface where we could connect to our BPM tool modeling software, which we could automatically transfer this BPM tool model into the process automation technology. 

I'd like to be able to do a deep dive into understanding how the engine works a little bit better, so that whenever I get called by an angry user about why something is taking so long, I'd like to be able to look through the logs like a developer would. I'm sure they could do a deep dive if I ever was able to engage them. The ability to do a deep dive would allow me to explain why something got hung up with in the process or explain how I can recover from a slow process.

In general terms, the current features fit our needs very well. Because the process design is something like a “blank page,” the product offers dozens of ways to achieve a specific goal. But, if I had to point out something to be improved, I'd mention the configuration. PAM connects to CA Embedded Entitlements Manager (EEM) for authentication/authorization, and to its database to store information about the process. These configurations can be changed only through the installation wizard. There's no way to update such information in the administration interface, which is, in my opinion, very bad.

Better reporting, because we have to true up our licenses every three months because we license on running concurrent processes, so we have to provide a report to CA every three months using their tool -- which never works.

From a usability point of view, I’d like to see improvement on their script editor so I don’t have to use two tools to generate scripts (right now I develop outside of the tool, bring it into the tool, tweak it, and roll out my workflow).

I don't use it enough to be able to speak to it right now. But, I'd say that, although it's drag-and-drop, there's a little bit of a learning curve to use it correctly.

• Regex
• Make some of the features more open source that way developers can have more flexibility.

• Auto-triggering of workflows based on machine learning (ML) and artificial intelligence (AI).
• Integration with web/Windows-based applications should be included.
• OCR capability should be added as a feature.
  

Now there are five different modules, but it’s hard to follow what happens in each location, so we’d like to see the solution more tightly integrated. The new release should be more integrated.

The solution engine is in Java. Regardless of developments, the solution will always be limited to what Java offers.

I think I speak for everyone: I’d like to have more best practices and samples in the documentation, and for it to be open for more product integrations.

Error handling on a per operator basis allowing native retries. For example the get email operator if there is a blip in the network and PAM fails to communicate with exchange it takes the error path when really I’d want it to keep retrying until communication is restored. It can be worked around but then it becomes time consuming and cumbersome to implement. It’s the same for other operators too. I’d like to see a search function on the operations tab when looking at past runs, it’s a right faff sometimes trying to find a previous run.