RPA has been on the news for some time. RPA, or Robotic Process Automation, represents a type of tool that includes several tools:

• A BPM

• A robot capable of reproducing human actions on software

• Very often possibilities around document scanning, and the ability to process mails

In practice, you can "record" human actions on software, such as a screen copy, a copy of a paper form, to drastically accelerate business processes. A certain number of people go so far as to say that it will revolutionize the work...I would not go until then because clearly, we can see that we could do without this kind of tool. Everything can be automated via APIs and a little code all around to make it simple. As an architect, I can only spontaneously cry out when I hear about it because architecturally, it could be likened to a wart. However, this answers two challenges that integration faces (and God knows integration is my chapel):

• ROI

• The availability of developers

Those are two brakes I've seen too much of. And in particular, the ROI. Yes, a BPM and an ESB would do a very good job of integrating business processes, but at what cost? With which AdHoc development reuse? Sometimes, the best architectural solution is the worst all because of the costs. So, let us not neglect this choice even if with more budget we would do otherwise.