Robotic Process Automation and Artificial Intelligence in Industry 4.0

Weird article. Apparently Robotic Process Automation is reducing repetitive tasks.

80+ years into the Information Age. And we're still confused.

Automation is not digitization. Most everyone conflates the two.

Automation is the removal of human judgement. This generally means simplification. All those biz pop terms like "business process reengineering".

Digitization is using computers to do tasks previously performed by humans.

IT (data processing) projects that set out to replace legacy systems, whether paper-based or prior software, generally fail. Because they do not first simplify the work being done. But rather attempt to fork lift current work onto a new media.

In other words, digitization without prior automation.

--

As for the selection of tools listed for Robotic Process Automation, they're fine mitigations.

When one cannot or will not first simplify, the minions are reduced to relying on shims and screen scrapping.

Ever more, I think most grand efforts like simplification, understanding, semantics, ontologies, and data interchange are quixotic. Just log everything and let future minions ferret out whatever interesting bits they might need.

In other words, enable future software systems archaeologists and forensics teams to better mine the data, and don't try to code for the ages.

Well, if you had AI that really was intelligent, it would be cool to have AI agents figure out your repetitive tasks as well as the structure of data in some giant IT app, and then turn this into a better, streamlined and automatable/automated App. The only problem is that AI is nowhere near that now and plausibly won't be there for a while.

But that's just wrong. Most RPA Software (like UiPath) uses selectors for web and desktop automation. You can use static or dynamic selectors, anchors and lots of other techniques to tell the bot where to click/type.

It's very dynamic and a well written bot will work even after changes to the website or software. Also, pure RPA is usually only used on legacy software without any other APIs, so they rarely change.

Source: RPA/IPA Dev

> First, get rid of software that holds your data hostage and cannot be controlled via APIs before you start thinking about “AI”.

The paper is talking about using AI to improve screen-scraping performance for RPA frameworks, which is definitely a valid use case.

As nice as it would be to always replace legacy systems with APIs, sometimes you have a job to do and need to get it done! If you are hired to build something and it needs to integrate with some windows application which can only be accessed via Citrix which you are not able to touch, then these tools come into their own (not as the preferred architectural choice, but as the only remaining practical option).

"get rid of software that holds your data hostage" - that can easily take an order of magnitude more money and time to do than implementing a screen scraping solution as an integration workaround.

I've seen my share of software like that. "Getting rid" of it may be a valuable thing in some cases, but it's not something you do "first", it's a strategic decision with an organizational transformation that will definitely take years to suceed and may be fail or be cancelled; it's something you might start but you may as well move on to a different position before it's finished. On the other hand, RPA is something you may actually achieve yourself in a reasonable timeframe.