RPA implementation roadmap
It’s important to understand that every RPA project is different. Each RPA implementation is shaped by the individual characteristics of an organization – including (but not limited to) budgets, technical capabilities, legacy techstacks, RPA vendor preference, executive buy-in and change management culture.
That said, there are key considerations that need to be addressed if the RPA journey is to be a successful one, offering a strong and enduring return on investment.
Discovery Phase
First things first, be 100% clear on RPA implementation goals – what you hope to achieve and what will justify the expense of deploying RPA solutions. Defined goals and success metrics matter. Though they may evolve over the course of the RPA adoption process, these considerations must be the project’s guiding principles.
Pinpoint those tasks and processes within the organization that are suitable for RPA automation, based on two key criteria:
- Ensure they are the correct type of task for RPA software – these tend to be rules-based, high volume, and repetitive tasks that are potentially prone to human error or involve handling large volumes of data.
- Analyze which workflows would benefit most from cost and time savings.
These considerations are crucial. As many as 30-50% of initial robotic process automation (RPA) initiatives fail because businesses target the wrong processes, automate inefficient processes, or fail to consider what happens to processes post-automation.
Planning
Determining the scope of the RPA implementation process should include research into indicative costs, timelines and required resources. This might have both internal and external components. Resource investment will be determined by the type of RPA implementation you opt for (we’ll get to that), so it may be worth talking to a range of RPA vendors.
Consult internally on project goals and scope – not only to get a view of likely team resource commitment but also spark stakeholder buy-in. Consider all the teams that could be impacted by RPA technology deployment and get them involved early, starting with IT. Even if an RPA vendor characterizes their solution as ‘no code’, IT professionals will be at the core of any integration with legacy systems.
Beyond IT, seek stakeholder support from across the business, from operations to HR and finance. This is the beginning of your project team that will maximize output and minimize disruption from RPA implementation.
All of this planning and scoping is essential if executive-level sponsorship of the RPA deployment is to be achieved. Without this support, there’s no RPA implementation.
Selecting your solution
There are a range of different models for implementing robotic process automation. The first of these is in-house RPA development, where an internal team develops and manages the software robots. This approach tends to incur lower long-term costs while offering maximum control, customization and internal expertise. But (and it’s a big but), the upfront investment and requirement for technical expertise is high. Realistically this model is most appropriate for digital businesses with specialist technical skills.
In the majority of cases, organizations will work with RPA vendors according to one of the following models (or occasionally a hybrid of different models when using multiple vendors):
- RPA Vendor Partnership: The vendor builds and deploys RPA bots based on clearly defined organizational needs. This approach offers easy adoption (and therefore quick implementation) and allows the organization to draw on vendor expertise. Customization options, costs, and potential for vendor lock-in (not all RPA solutions play well with others) will vary according to the individual solution.
- RPA as a Service (RPAaaS): RPAaaS is a cloud-based model for using RPA tools where the RPA software and infrastructure is hosted and managed by a third-party provider rather than building and maintaining it within the organization’s techstack. This vendor-managed option offers rapid deployment, low upfront investment, and easy scalability. On the flip side, there is increased vendor dependence, reduced control and customization, and reliance on third parties to manage security and RPA infrastructure.
Within each model, there are many RPA software vendors providing a range of different solutions. Take some time to compare the features, integrations, level of internal resource required, level of external support provided, scalability, and pricing.
The key is to find the RPA solution most suited to achieving your implementation goals within the constraints of budgets, time frames and in-house technical capability.
Building Better Bots
Whatever pre-built out-of-the-box integrations your selected RPA solution provides, successful implementation relies on close collaboration with your vendor on bot design and deployment. RPA development needs to be rooted in business best practices and bot behaviors optimized to replicate (and improve) human workflows.
Through this collaboration, RPA bot design will be shaped to incorporate all required application interactions, data transfers, error handling, logging, notifications and triggers.
Training and Change Management
One of the biggest RPA implementation challenges is cultural rather than technical – winning the hearts and minds of the workforce around the benefits of process automation. Successful implementation requires teams to be open minded as a minimum, and ideally to embrace the new working practices.
But deployment of RPA technology necessitates changes to the roles and responsibilities of employees, and there may be resistance from individuals fearing job losses. So it’s vital to be transparent in addressing such concerns, to emphasize the positive impact to the individual, and to provide comprehensive training where required.
Deploying and Monitoring
Test, test, and test again. Work with your RPA vendor to pilot limited bot deployment in discrete areas and monitor bot activity and performance data closely. The integration of software robots can then be ramped up gradually and safely – using proper access controls, encryption, and security monitoring. Any RPA developer will tell you that this initial deployment is not the end of the story: using performance data and user feedback bot activity can be refined over time.
RPA Governance
Like every other enterprise IT asset, RPA software requires clear, centralized oversight and governance. This will entail creation of policies for RPA bot access, data security, change controls, and even disaster recovery. Similarly, controlled, monitored procedures need to be established for robot maintenance, issue tracking and upgrades.
But RPA governance also entails careful monitoring of changes to tasks, processes or workflows that intersect with automated processes. An RPA tool won’t detect or react to these changes without instruction – but the changes could compromise, or entirely derail the process automation.
Measuring, Optimizing and Expanding
No RPA use case is a one-and-done exercise. It’s iterative, evolutionary. Therefore it’s vital to measure the impact of process automation against expectations and project success metrics in order to refine RPA software performance. It’s also important to track any unintended consequences of the RPA implementation (whether positive or negative), and to identify potential new use cases for future automation.
