1. Peter Schooff
  2. Sherlock Holmes
  3. BPM Discussions
  4. Tuesday, 05 February 2019
  5.  Subscribe via email
There have been more than a few mentions of the bubble bursting with RPA in the past few months. What do you think?
Accepted Answer Pending Moderation
Worst case scenario is that it will plateau. Large enterprises will have realized that robots subscription is one very foreseeable recurring cost, while robot retraining on every single legacy change is one very unforeseeable (but almost recurring) cost. They will either stop completely changing anything in their legacy systems (may God have mercy on their souls) or will wean off the Kool Aid.

Most likely scenario is that it will continue to grow at a sustained (but not frenzied) pace for the next 1-2 years to a solid enterprise valuation, and then we will see the large players exit to either very big IT enterprise juggernauts who just want a bigger piece of the IT spend wallets of their already captive customer base (may God have mercy on their souls) or via IPOs to retail investors (may God have mercy on their souls too!).

Best case scenario is that RPA players, after the AI marketing play will fizzle out, will expand horizontally into other forms of automation, in order to capture end-to-end processes as well (may God have mercy on our souls!).
CEO, Co-founder, Profluo
Accepted Answer Pending Moderation
Generally agree with Bogdan's comments, who is very gracious in extending mercy on behalf of the divine. :D

For the simple processing stuff that is 1st Gen RPA, by which I mean the fairly algorithmic stuff that is reducible to something like a nested set of if-thens in on-screen actions/reactions, then yeah, worst case scenario is it plateaus as it is adopted quickly and widely to extend life of existing UI-based apps. BTW, retraining is a modest net-increment to the functional testing already done for any release, where you take the modified UI elements and behaviors after confirming the correct functionality as the new basis for teaching RPA what to do, so it's not that big a deal, IMO.

I do agree that an industry shakeout is inevitable, and it won't be pretty (because it never is). However, IMO, the future value prop for RPA is in extending how much of the work it can take on, moving from the more algorithmic, in which next step can be programmed, to the more cognition-based, in which AI determines next step. We've basically started with RPA as an automated programming of programs (or at least programmed behaviors), so as we turn more and more over to RPA and AI to (in effect) program for us, the more and more it can do.

And, yes, when the singularity hits, we will need all of the mercy the divine (or our robot overlords) will grant us. ;)
"BTW, retraining is a modest net-increment to the functional testing already done for any release"

Hi Lloyd, I agree that UI changes do not require extensive retraining.

My comment was related to back-end legacy changes - like when data is moved to another system, or has a decrease in quality due to some validation procedures breaking down, or some re-factoring of data taxonomies that change task logic and role-based credentials etc. I have seen a couple of these where people realize (and calculate) it's almost a re-implementation.

No wonder big consulting companies are excited about RPA. Nobody would get that excited only for the initial implementation revenue :D
  1. Lloyd Dugan
  2. 2 weeks ago
  3. #5925
Agreed. I do with, though, that there was some evidence on the ROI of extending utility of a UI-based app through use of RPA (and a dash or two of AI) over simply refactoring the app to permit more standard STP approaches (as "headless" BPM, as I call it, as opposed to robotic BPM, which is still a "head" though it is a virtual robotic one). I know that this is point of contention because I had encountered it with the Case Mgmt App Dev Team passive-aggressively resisting RPA because they wanted to exactly that, and were never quite persuaded by arguments that such would prove more expensive and longer to achieve than we had time to do.
Accepted Answer Pending Moderation
I mostly agree with Bogdan Nafornita : I wouldn't expect the bubble to burst, but a slowdown in RPA adoption and growing.

The main thing here is that RPA doesn't seem a long-term solution for system integrations. It is a great tool in the short term. Of course, it can produce immediate results, having the robots working 24x7x365 in boring tasks like moving data from one app to another at a very affordable cost per transaction.

In the long run, when you have more and more robots, and applications change (and they really change over time), it could be very hard to manage. Some applications integrated through screen reading, other through a win32 app with form fields, and other through web interfaces. And dozens of robots reading those interfaces. If they change, it could be a nightmare to reconfigure all robots again and - most important - testing that they are properly working.

As previously noted in this forum, RPA is a great instrument for some specific needs. It should be integrated into more complex business processes to solve certain situations. But in my opinion, it won't continue growing at last year's rate.
CEO at Flokzu Cloud BPM Suite
Jim Sinur
Blog Writer
Accepted Answer Pending Moderation
I don't think RPA is a bubble as a whole as there is much more for RPA and automation, in general, to do to complete work for organizations. You can read about the top five on ramps to RPA in the link below. RPA will step up to these five opportunities and take on more complex tasks as it partners with BPM/workflow, process mining, data science, smart cloud platforms and AI vendors. There will be eventual consolidation and some individual RPA vendors and their investors will lose, but an overall bubble is unlikely. If RPA vendors do not partner or merge with other players and only pursue narrow task management in processes only (probability 10%), then maybe a market bubble. I am already seeing evidence of cooperation and buyouts.
  1. https://jimsinur.blogspot.com/2019/02/the-top-five-on-ramps-for-rpa.html
Accepted Answer Pending Moderation
The current miserable and unfortunate state of digital transformation will keep RPA well alive.

Accepted Answer Pending Moderation
It depends if RPA vendors overpromise. The P should stand for Procedure not Process, since RPA does not automate processes.Even BPM has had difficulty with this.
I have run into 2 very large transformation projects where the P was oversold and the projects stalled. In one case BPM was bought in to solve the situation but then the customer had no consolidated definition/picture of what they were doing.
The fact that BP founders sold stock while pitching huge growth potential may imply they envisaged a burst.
@Anthony. I find your "The P should stand for Procedure not Process" very interesting.

I suspect most of the players at this forum think of "process" as a mapped, compiled and rolled-out sequence of tasks that is capable of providing background orchestration in a run-time workflow/workload management environment.

A "procedure" can be a "process" but it can also be the result of observing,once, a human carrying out a sequence of tasks that an RPA system observes and is then able to replicate.

RPA 2.0 adds AI to RPA, but, in my view, we don't need to be in a rush to go there because much of the potential of RPA is yet untapped.

Suppose I want to build a Kbase comprising feeds via e-mail of, say, terrorist incidents - if a source can be found that pushes out individual event notices, we have the choice in the Kbase app of building a parser that periodically reaches out to read incoming e-mails, extracts the content, then, back at the Kbase, adds a new data point per e-mail message at a form that features a field capable of hosting such content.

Alternatively, with RPA, it should be possible to simply have the RPA system observe a user sequence, with the result that the RPA app then periodically scans incoming e-mails looking for terrorist event notices, scanning/extracting the content and appending new data points (i.e. data nodes) to a container or bucket in the Kbase. Here, I am saving programmer time.

On the other hand, a one-time copy-paste of names, addresses, phones etc data from a text file with the objective of appending new data points saves user time.

Lots of opportunities for increasing efficiency plus improvements in effectiveness as a result of consistency in the performance of iterations.
Accepted Answer Pending Moderation
I do believe the hype around RPA will end sooner or later, when customers that have been oversold by RPA suppliers reach the conclusion that while RPA is nice addition to their workforce, it is nothing more than the replacement of a human actor by an automated (programmed) one in a specific task in a process. [by the way, a process to me is a described sequence of activities that transform an input into a more valuable output and is nothing something that can be 1-on-1 automated, a more detailed level of modeling would be required for that].

I guess RPA will follow the same path as most tools in the bigger BPM picture do. Hype at first ("they are going to massively disrupt your business";), despair next ("oops, this didn't bring the estimated benefits after or at all";) and commoditization in the end ("a lot of things are done via (ro)bots";). My take is it that we are on the brink between hype and despair.
BPM is all about mindset first and toolset later....much later
Accepted Answer Pending Moderation
Agree it does seem that RPA has been over hyped sounding good but very limited in what it actually does; its not automating processes. So either it will collapse with disappointed buyers OR it needs to move to covering all the automation possibilities within business processes such as delivery of Adaptive UIs, creation of letters & e-mails, notifications, escalations, reports etc. And of course calling on legacy/external data as required on the UI. This puts into a RPA Platform which needs to cover all the underlying business process logic requirements. However reality RPA as it stands RIP.
  • Page :
  • 1

There are no replies made for this post yet.
However, you are not allowed to reply to this post.