Business leaders have embraced Robotic Process Automation (RPA) technology from the likes of Automation Anywhere, UiPath, Blue Prism, AssistEdge and Pega Robotics as a powerful tool to improve productivity and take cost out of the organization.  Shared Services and BPO Centers were early adopters of this technology as they have an abundance of repetitive deterministic tasks that can be easily automated with RPA solutions.  With proven success in the core shared services functions (Finance, Order Management, HR, Procurement, IT, Customer Service), interest in RPA has extended to a wide variety of business areas.  But with success comes a variety of challenges.

The Challenges with Automation

As I travel around the world talking to business leaders on the topic of automation, there are six key challenges that just about every large scale enterprise is grappling with.

Fig 1: Six Key Challenges with Implementing Automation

  1. Automation Strategy – What are the vision, charter, roadmap and measures of success for automation? How to identify attractive automation candidates?  How to prioritize and justify automation investments with a compelling business case?
  2. Technology Strategy – What is the right automation technology solution?  How and when to embrace more advanced automation technologies (assisted automation, chat-bots, OCR, artificial intelligence/machine learning)?  How to build a scalable automation architecture?
  3. Governance – What types of governance bodies are required? How should automation investments be funded and who should have what decision rights?  What are the new roles and responsibilities?
  4. People Enablement – What kinds of skill sets need to be built in order to support automation?  What should be the Change Management strategy?  How do I manage the people impacts of automation / reskill people for new ways of work?
  5. Operating Model – How to organize around RPA?  Should RPA be a business or IT solution?  What kinds of controls need to be put in place around the technology?
  6. BOT Development & Support – What are the right development methodology, tools and standards for automation?  What do we design for reusability and manage robots as an asset?  How to manage and monitor an RPA environment?

Because the barriers to entry for RPA technology are so low, individual business units can pilot RPA technology, or even build a RPA delivery capability, with little or no involvement from IT.  Issues emerge when:

  • Multiple business units embrace competing RPA technologies
  • The volume of robots reaches a scale that the business can no longer support
  • Changes to the underlying systems break existing automations that IT may or may not be aware of

Will a Center of Excellence Improve Automation Success?

To address these challenges, large enterprises often conclude that they need to establish a Center of Excellence (COE) around automation, more broadly, or RPA, specifically.  I recommend that enterprises consider automation technology across the spectrum of work, not just RPA technology, when establishing an Automation Center of Excellence.

The internal organizational politics around establishing an Automation Center of Excellence can be fraught with conflict as multiple parties jockey for position around the next hot technology.  By focusing on a few universal design principles, it is possible to design an efficient operating model:

  • Avoid Duplication of Effort and Expense – No one wants to waste firm resources by reinventing the wheel. A good COE design will centralize those functions/capabilities that can be done once and re-used across the enterprise
  • Focus on Time to Market – RPA is all about rapid deployment of non-invasive solutions, so it is critical that RPA solutions not get bogged down with an overly burdensome and time-consuming process or structure that increases costs
  • Business Agility – In order for process integrity to be maintained, the business must feel ownership of all automations that are created by deciding what and when to automate
  • Leverage Economies of Scale Where Possible – A minimum viable scale exists for creating economically-viable automations.  A free-for-all where RPA solutions are developed across the enterprise is probably not the right answer
  • Protect Against Risks – There are a number of risks inherent in automation solutions that anyone who develops them should be aware of and mitigate against

Choosing the Right Automation COE Operating Model

The basic structure for an Automation Center of Excellence can be divided into two parts.

Fig 2: Operating Model Components for an Automation Center of Excellence

Automation Center of Excellence – The core of the Automation COE are those things that are done once and then applied over and over again.  The Automation COE should establish:

  • Automation Strategy – Define the framework for how automation opportunities are identified and prioritized. Create the overall automation COE business case, as well as provide a business case framework that can be applied to each automation opportunity.  Define the program metrics and measure value realization
  • Technology Solution – Select the automation technology, define tools and standards, manage software licensing, manage vendor relationships, provide the hosting solution, define security standards and monitor bots in production
  • Governance – Define governance bodies, define decision-making rights, define funding mechanism, define stage gate criteria and define standards and controls
  • People Enablement – Provide technical tool training, provide the Organizational Change Management (OCM) methodology, tools and templates
  • Operating Model – Define the service catalog, provide deliverables templates, define development standards and maintain bot asset inventory

Scalable Automation Execution Engine – The recurring activity of defining, designing and delivering automations is horizontally scalable and can either be centralized as part of the Center of Excellence or federated out to multiple delivery centers.  The scalable automation execution engines should be responsible for:

  • Demand Management – The identification, prioritization and justification of specific automation solutions
  • BOT Development – The design, build, test, deployment and stabilization of software robots
  • BOT Support – Ongoing support for software robots in the production environment, including exception handling
  • Organizational Change Management – The organizational change management, resource redeployment, separations, severance and training associated with process automation

Based on the capabilities described above, there are three basic options for operating the two components of an Automation COE:

Centralized Model – In a Centralized Model, all the functions of a COE are performed by a single joint team of personnel from both the business and IT.  This model works particularly well when:

  • An enterprise is just establishing its automation capabilities and there isn’t any existing capacity elsewhere in the organization
  • A strong shared services culture and charge-back mechanism already exist
  • The enterprise is highly-centralized
  • The automation opportunity is modest (less than a thousand addressable FTEs)

The benefits of a centralized model include:

  • Easier to enforce processes, policies and standards
  • More effective knowledge re-use
  • Easier to achieve economies of scale

The drawbacks of a centralized model are:

  • If demand for automation exceeds capacity, resource contention will result in some groups’ automations being deprioritized
  • In a sprawling global organization, some parts of the business may go off on their own and duplicate effort if they aren’t aware of the capabilities within the COE. Automation evangelism becomes critical in large enterprises

Decentralized Model – In the decentralized model, all the functions of a COE are replicated for each business unit.  By its very nature, you aren’t really creating a center of excellence; you are creating multiple communities of practice (so, “COE” is probably not the right name for this structure).  Information can be shared between Communities of Practice to encourage standards and leverage common technology, but this model is ill-advised for a number of reasons:

  • It creates duplication of effort and headcount, resulting in a higher cost for automation
  • Even the best-intended people are prone to standard drift when responsibilities are duplicated across organizations

Federated Model – In a federated model, the Automation Center of Excellence capabilities are centralized within a single small group, but the Scalable Automation Execution Engine capabilities are federated out amongst business units, or as many groups as required.  The Automation Center of Excellence may also contain an Automation Execution Engine that provides automation delivery capabilities to business units that lack sufficient scale to justify their own stand-alone Automation Execution Engine.  The benefits of this model include:

  • Easier to enforce processes, policies and standards
  • More effective knowledge re-use
  • Easier to achieve economies of scale
  • Each business unit is able to proceed independently, avoiding the challenge of competing priorities
  • The business units will feel a stronger sense of ownership and engagement over the automation journey, as they are accountable for their own execution engine

The federated model works best when there is a large, globally-distributed enterprise with highly-decentralized decision making.

In conclusion, an Automation Center of Excellence can improve an organization’s automation success.  The decision to embrace either a Centralized or Federated model for operating the Automation COE needs to be driven by the organization’s unique circumstances.  As long as the operating model is structured properly with clearly defined roles and responsibilities, either model can be used to achieve your automation objectives.

This is the third post in our series on Automation.

Read Part 1: Automation Technology Across the Spectrum of Work

Read Part 2: Picking the Right Automation Tool for Your Job

Joshua Biggins

Joshua Biggins

Partner – Enterprise Strategy & Architecture

Joshua Biggins is a partner with Infosys Consulting where he leads the Enterprise Strategy & Architecture practice for a number of industry verticals.  For the last 22 years he has focused on helping clients leverage technology to transform business models and unlock value.  His experience is focused on the most pressing issues on the CIO agenda, including AI and automation, IT cost reduction, application portfolio rationalization, managed services transformation and technology modernization.