RPA : only 4% of companies have already deployed more than 50 robots nowadays

RPA (Robotic Process Automation) is a fashionable technology which is demonstrating its potentialities inside the companies that have launched initiatives about it. However, a while ago, I “felt” on a publication already announcing that “RPA is dead”. Beyond the voluntarily provocative title, it is necessary to distinguish things.

RPA (Robotic Process Automation) is a fashionable technology which is demonstrating its potentialities inside the companies that have launched initiatives about it. However, a while ago, I “felt” on a publication already announcing that “RPA is dead” (1). Beyond the voluntarily provocative title, it is necessary to distinguish things. In order to give again sense, business and economic interest for the company and its collaboratives inside an integrated system, it is necessary to consider RPA and to set up it through a holistic approach.

Currently, companies are mostly acquiring and enforcing robotization of process technologies. “Although a big number of companies got started in RPA, only 4% of them have deployed more than 50 robots nowadays” (2) even if they present themselves as already well-advanced into procedures. This state has been very well illustrated by Sébastien Dumas (Synertrade)’s conference during the Paris’ CNA during May 2019. Among around twenty participants, only one had already deployed (inside his organization) robots in an industrial way (more than 30 robots) and identified limits of RPA when it is implemented alone.

RPA’s limits

One of the main identified limits of RPA into our current works sticks to the consideration of the implementation of this technology rather than its intrinsic capacities. “Do not forget that all of these new technologies are a way and not an end in itself” (3)

In order to widely deploy RPA, approaches based on the one hand on processes’ cutting and on the other hand ones aiming to robotize all processes are implemented. Indeed, by considering individually processes or tasks, it becomes more easier to apprehend global complexity, to deal with divergent events (by categorizing or excluding them) and to potentially implement a big number of robots.

Doing this, problem to deal with is delimited but sometimes considered as a unitary element. What works in other domains (such as some modelization or numeric analysis methods) does not operates in a sufficient way in the business world and in the vision of integrated processes into department (purchasing, accounting, management control, HR, etc.), or organization (regional, national or global).

Thus, some « RPA early adopters” lose sight of potentiality of RPA, such as the improvement of realization of an integrated process and privilege a too-fragmented vision of process. Doing this, they optimize one or several tasks of the process but do not improve the process in general. Existing “bottleneck” are for example not processed.

From this statement, RPA’s implementation approach should be multiple:

  1. To propose an oriented-to-integrated-processes approach
  2. To propose a vision of RPA integrated to the company’s digitalization strategy and to the company’s strategy
  3. To propose a technologic toolbox that permits rational combination and enhancement of RPA through Analytics and AI.

Process approach : a coherent to robotization approach has to be preferred instead of a massive and generalized approach

As part of interventions leaded around robotization, it is necessary to exercise discernment regarding the application and implementation of technology serving the company.

Even if it is sometimes difficult to entirely robotize process, it is necessary to incorporate a dose of good sense in the approach of robotization’s perimeter. If it is not possible to robotize a process, it is not necessarily desirable to robotize an isolated task. It would indeed bring little of added value, because the improvement of a task’s efficiency does not necessarily improve the process in its globality. Automation of a task can only move bottleneck into the process’ development. A vision under the sub-processes’ angle seems to be a median approach which not entirely decontextualizes robotic elements and permits thereafter a more global vision.

Most important starting point is to not tend to robotize all processes from the time necessary and sufficient conditions are not fulfilled. Standardization, known errors, structured inputs, repeatability, processes based on management rules, etc. These choice criteria have to be considered in order to measure eligibility to robotization and to select in a pragmatic way robots’ implementation.

Structured approach of RPA at the company’s level : strategy and support of the human in order to favor the adoption and to lift fears.

As we mentioned it above, all-levels robotization should not be implemented at all costs. It would accompany itself with a reject of technology by collaborators and would be in fine counterproductive.

RPA, as well as other initiatives into companies should be considered in a global way and be structured in order to integrate itself :

  • Into company’s digital roadmap. It will permit to select more-adapted tools, to embark DSI into procedure from the beginning of it and to propose a middle to long term vision.
  • Into global company’s strategy. This integration will permit to associate technology’s implementation with organizations and persons’ support while their adoption and deployment of RPA, in order to permit a transformation towards a hybrid taskforce (humans and robots). Robots and humans are indeed completing themselves without necessarily letting the first ones replacing the second ones.

Technologic toolbox: from the robotization to the integrated and intelligent automation.

Business issues cannot be processed and resolved thanks to a unique technology. ERP have not resolved all company’s ailments. It is thus necessary to consider RPA as a technologic brick into a toolbox that permits to answer to company’s issues. Artificial Intelligence, process mining, Analytics, data visualization tools, NLP engines but also ERP and Purchasing SI (knowing that it is not an exhaustive list) should be integrated into the technological vision in order to toggle from robotization to “Integrated and Intelligent automation”.

Recent evolution of technologies turns this integration in an easier one than past and multiply possibilities, functional coverage or also contribution of values for process automation.

As a synthesis and in order to consider RPA as a business accelerator

  1. RPA should be part of a strategic and extended vision. Once POC step is finished, a robotization roadmap has to be built and operationally declined.
  2. RPA should lean on a reinforced Change Management program in order to make possible transition to a new work organization (hybrid) and to secure human into the company.
  3. RPA should not be considered alone, but increased with other innovating technologies and existing tools for the company.

Benjamin Verdier

Références :

(1) RPA is dead. Long live Integrated Automation Platforms

(2) RPA : robots are waiting for

(3) AI and Blockchain applicated to Purchasing cannot replace fundamentals