Why Do Bots Fail to Scale Across the Enterprise?
The interest in RPA has skyrocketed, and company leaders are challenging their teams to find out more about the technology and its associated benefits. With the increased interest in RPA, we have seen a significant uptick in teams testing the RPA waters by starting Bot development and implementation pilots. What we have also found is that teams are struggling to move beyond the pilots due to some fundamental errors made during RPA Program Setup and Execution and Bot Development and Implementation.
RPA Program Setup and Execution
What we find is that there is a lack of an RPA enterprise strategy and foundation, and a lack of understanding about RPA, solution capabilities and where to focus efforts.
Strategy and Foundation
Multiple functions throughout organizations have stood up disparate RPA solutions in silos and started creating bots without engaging central enterprise teams or IT (Beers). Many RPA deployments have become a free-for-all, which has created an unmanageable proliferation of RPA solutions and bots. This has led to dramatically increased disorganization, costs, and difficulties in quantifying RPA benefits (Moore)
Solution Capabilities and Focus
Leaders and business process owners are excited about the possibilities for RPA. However, they lack understanding of the solution’s technical complexities and capabilities which need to be taken into consideration when determining which processes / activities / tasks to automate (Tornbohm). When decisions are made regarding what to focus on, there appears to be a tendency to take on too much at one time by trying to automate an end-to-end process or by trying to automate a workflow that is fundamentally broken (Abbattista). In these cases, bot development complexity, time to delivery and costs increase way above what should be a 2-3 week effort.
Bot Development and Implementation
We have noticed that there is a tendency to develop bots without expert RPA professional guidance, and forego establishing necessary RPA management, support and change management.
Expert Professional Guidance
Getting started with RPA is relatively easy. However, gaining access to expert RPA architects and developers is difficult given the newness of the technology (Patel). Given these dynamics, organizations are randomly selecting teams to design, develop and operate bots. The challenge becomes that these team members tend to do so in an ad-hoc inefficient manner (Abbattista). RPA scaling takes planning and development by an appropriate, well-versed, and helpful RPA team. When expert RPA architects and developers are not engaged, the result tends to be lower-performing, error prone, non-scalable and costly bots.
Management, Support and Change Management
RPA pilots are successfully demonstrating the art of the possible, which is driving enthusiasm about the technology (Moore) and what it can do for organizations. However, RPA programs quickly run into significant challenges when bots become unstable or broken, are used for limited activities or are quickly abandoned. The reason for this is that:
- RPA bots are often thought of as “set-and-forget”, when they actually need to be managed (operationally) and maintained (technically) (Muraleedharan);
- The IT organization is often not engaged to help provide the ongoing support for the bots (Abbattista) so they are “not properly embedded into core systems” (Grung-Olsen) or ITSM processes; and
- Stakeholders want to accomplish too much too soon including enhancements in automation capabilities to perform Machine Learning and AI tasks
- Key stakeholders and employees may not be engaged early and buy-in is not achieved which causes resistance, resentment and fear of job loss within employees (Abbattista) (Moore) (Patel).
Creating and Executing Enterprise Scalable Bots
To ensure RPA efforts result in bots that are scalable across the enterprise, an RPA team needs to setup and execute a robust RPA Program and put in place a structured bot development and implementation process. To assist RPA teams on their RPA journey, we lay out key tasks that should be completed to help the team build enterprise scalable bots.
Robust RPA Program
To setup and execute a robust RPA program, the RPA team needs to establish an automation strategy with a governance structure and a detailed roadmap. The team also needs to obtain buy-in from cross organizational stakeholders on “Task Bots” to be created.
Strategy, Governance and Roadmap
- Create an RPA strategy, develop an operating model and governance structure, capture and rank RPA opportunities, build the business case (Muraleedharan), and shape a roadmap (Thronbohm);
- Ensure that the RPA solution is properly integrated into IT’s worldview and strategy (Muraleedharan) in terms of security, reliability, scalability, continuity and fault tolerance;
- Work with system administrators, compliance, HR, cybersecurity etc. to agree on how the Digital Workforce will gain access to systems involved in the target opportunities; and
- Engage an RPA business partner (Abbattista) to obtain advice on the direction of the RPA program, access to skilled RPA developers and training to up-skill the in-house team.
- Create a view of the in-focus enterprise processes and clearly articulate which tasks (Trefler) would be automated and which team members would be impacted;
- Work with Business Process Owners / Subject Mater Experts to identify example use cases and then video record the keystrokes / mouse clicks / swivel chair activities;
- Take time to think through how the tasks would need to be reconfigured so that each bot can perform agreed work with minimal human intervention;
- Build and record proof-of-concept bots to be shown to key stakeholders; and
- Prepare for and socialize the bots with key stakeholders noting requested changes, challenges that need to be overcome, and gaining buy-in on the changes (Abbattista).
Structured Development and Implementation
To put in place a structured bot development and implementation process, the RPA team needs to build resiliency and scalability into bots and formalize bot deployment, management, support and control.
Resiliency and Scalability
- Form a bot development team that consist of an experienced bot developer who knows the RPA solution tips and tricks, a business analyst, tester(s) and a change manager;
- Document requirements, especially noting exceptions and how errors will be handled;
- Plan for development, testing, piloting and deploying the bots;
- Develop the bot in a modular fashion (Moore), reusing existing bots (Fuochi), capturing objects in a way that mitigates cross system challenges, and including error handling; and
- Run agile pilots to test the bots in the real world, learn at speed (Muraleedharan) and harden the bots in preparation for scaling the bots across the enterprise.
Deployment, Management, Support and Control
- Plan for either enterprise wide or incremental deployments, which includes development and execution of organization change management, communications and training plans;
- Outline and agree bot runner and orchestrator roles and responsibilities, and identify individuals from across the organization to fill the roles;
- Leverage the RPA solution capabilities to remotely model, monitor, control, schedule, and execute (Hankiewicz) the enterprise bots;
- Scan the organization for existing support models, and build a support model that sets out local operations and centralized function responsibilities (Abbattista); and
- Execute a change control strategy that ensures that bots are tuned, processes are kept up to date, systems are aligned, and human workers are informed (Muraleedharan).
Successfully accomplishing the above actions requires committed internal and external resources focused on getting the RPA deployment right (PEX), a willingness to engage with and follow existing organizational demand management, project delivery, risk management and product management frameworks. There also needs to be a clear focus on creating attended bots (Trefler) that fix specific activity / task challenges and engagement with and buy-in from cross organizational stakeholders. Lastly, the RPA team needs to ensure that there is dedication to executing a structured agile approach to bot development, implementation, deployment and management.