What To Look For in a Robotic Operations Center
Implementing RPA is great, but not the end of the story. You may need a Robotic Operations Center. Read this article to find out more.
Join the DZone community and get the full member experience.Join For Free
Robotic Process Automation (RPA) is the magic wand for organizations looking to enhance everyday operational efficiencies. Besides drastically reducing time spent on low-value tasks, eliminating errors, and improving accuracy, RPA helps enterprises meet business goals and drive value. Interestingly, this journey towards RPA is incomplete without a strong Robotic Operations Center (ROC) strategy that helps organizations address automation challenges and amplify RPA adoption.
This article is a tutorial on understanding the criticality of speedy RPA deployment, as well as the importance of dedicated bot monitoring and maintenance.
ROC: A Quick Introduction
A Robotic Operations Center (ROC) is a 24/7 maintenance and support program that enables enterprises to manage automation deployments. The Managed ROC utilizes a broad set of tools and utilities that allow the ROC team to support and scale RPA implementation at a lower cost while leveraging onshore resources. These tools and utilities include:
- Automated Deployment Utilities
- Automated Provisioning Scripts
- Alerting and Monitoring Dashboards
- L1, L2, and Automation Infrastructure Support
An RPA lifecycle starts at the discovery stage to identify the process, process categorization, and validation, and ends with bot development. ROC is the next stage of RPA. Deploying automation capabilities to the ROC to support production, monitoring, and optimizing bots using a skilled team and providing continuous support is where the ROC really pays off.
A comprehensive process supports DevOps, cloud infrastructure monitoring, and network architecture.
Deploy bots into production to review the code and check for compliance while you do so. Also, adhere to multiple security scans. Depending on the requirements, run tests based on the playbook deploying the bots into the production environment, adhering to the configuration settings as required. When moving from development to QA, the configuration needs to be updated for corresponding environments, and smoke tests need to be executed to make sure the parts are working in production.
Bot monitoring is an important part of the ROC process, where the team proactively monitors the bots for performance. Capacity management helps us study how the VMs are effectively used. With weekly or monthly reports to better optimize the bot infrastructure, the team troubleshoots issues based on the playbooks and then provides the Root Cause Analysis, leveraging these insights to address performance improvements, minor fixes, and version upgrades.
While the bot infrastructure is typically hosted on the Cloud, the actual system must run on your local environment, because it needs to connect to local desktop applications or other internal systems as required. Bot monitoring in this case will be managed in the cloud.
In this stage, no code will be tested. L1-L2 is deploying and monitoring RPA bots. If there is an issue, the operators identify the root cause and then notify the teams concerned (including the development team) so they can fix it.
L1 support is a help desk resolution service ticket that takes care of basic requests such as password reset. Issues that cannot be easily resolved by L1 are sent to L2, where agents have access to certain advanced configurations that can address complex issues. Requests beyond the scope of both L1 and L2 support are forwarded to L3.
L3 support is where you need to test the code. For example, in your product, maybe an element has been changed, maybe the text field, or maybe the name field is not yet in the same equation, and it won’t click through, mandating the need to add some more code.
P1, P2, P3, and P4 Issues
Critical client issues are classified under this category, P4 being the lowest and P1 being highly critical.
The team communicates these through email and other collaboration tools. Work within the client’s native environment. Our team members are added to the client’s ticketing system and tickets are assigned to them in that system.
A Comprehensive ROC Engagement Model
One popular engagement model is a mix of onshore and offshore ROC managers, a 24/7 service desk, and a core engineering team working in tandem with your automation team. The meeting cadence is a weekly touch-up, weekly calls, monthly calls, and quarterly calls for a comprehensive operational review as well as to address technical and strategic issues.
The engagement starts with an initial two- to three-week onboarding involving:
- Discovery session (two to three weeks): Includes understanding the current bot scenario, documentation checks, and process evaluation.
- Knowledge transfer (four to six weeks): Involves understanding bot complexity.
- Shadow support (four to eight weeks)
- Operational takeover
Follow the global support model with team members including infrastructure engineers, RPA support team, and pure RPA developers and architects working from different time zones. A dedicated vendor-client team is responsible for coordinating with product tool vendors.
Bot Monitoring and Metrics
While bot development is a focused activity on one bot, when it comes to support, ROC teams need to provide support to about 20-30 bots at one time. They need to clearly understand bot functionality, code nitty-gritty, and documentation. Bot monitoring also involves monitoring the on-premise tool or system. In the case of UiPath, we monitor the control room. If it is Automation Anywhere, we monitor the orchestrator or BP orchestrator, besides GitHub metrics. Apart from the tool, we also monitor business KPIs, including:
- Duration of the bot in action
- Records/Orders processed
- Average processing time
- Successful/Failed records rate
- Bot Full Time Equivalent (FTE)
- Time utilized by the bot
- % of NCP utilization
- Memory consumption
- Maximum efficiency in terms of VM
Get daily reports comprising pending incidents, escalated incidents, repeated call analysis, shift handover, and URL reports. This daily data is then consolidated into weekly and quarterly reports. The quarterly report helps measure the SLAs, how many tickets are addressed, and how the bots are performing, besides infrastructure status, and license utilization for SLAs.
Interactions With Tool Vendors
For niche issues with any of our ROC vendor partners, we communicate with the external vendors, data vendors, and other third-party services support partners for timely resolution. A dedicated backend technical team comprising solution architects and engineers works round the clock to provide continuous support to clients.
For example, if an automated scheduled bot supposed to run every 30 minutes stops doing so, our team will follow the playbook and check the cause of the malfunction. They will try to check the logs and see if there are any issues with any of the applications. Maybe a third-party application is down, or maybe network connectivity is down, or perhaps disk space is not running at its normal capacity. They’ll clean it up and then run the bot again. Regardless of the root cause, we’ll fix it.
For network connectivity or similar issues, the ROC may need to delegate it to the IT team. However, if it is something like disk space, a disk check issue, or a password change that is required, we'll coordinate with your team and take the appropriate action to resolve the issue.
Enable your organization to accelerate automation initiatives by providing end-to-end consulting and support. A team of automation experts deploy, monitor, and manage RPA projects including dedicated Robotic Operations Center services. Have a set of established guidelines around security and coding best practices that we follow as part of the development process and review the code before deployment to ensure there is no data leakage.
Opinions expressed by DZone contributors are their own.