So the consultants are gone, and have left you the responsibility of managing the bots they built and pushed into production?
Let me guess...
You weren’t consulted during the build?
There is little to no documentation?
Those who built it are gone now?
Business leaders are holding you responsible for bot support and maintenance?
If the above applies to you, don’t worry! You are not alone. IT departments across almost every industry are waking up to this new reality.
Automic is providing resources and content to help navigate this opaque landscape.
While the problem is multifaceted, and often requires bespoke solutioning dependant upon your industry or the various functions the bots “automate”, we’ll be releasing a series of the articles, created by our experts who’ve been there a few times, with insights that may be helpful in navigating your automation support journey.
One of the first things on the mind of the CTOs / CIOs / IT leads we work with is,
“How do I resource the ongoing bot monitoring and support required?”
So we’ll start there.
Resourcing as a function of bot volume
Allocating sufficient resources to support a scaling automation program will require determining a baseline level of support. In other words, determine what is needed to manage a single bot, and then extrapolation from there. To do this we’ll outline three main roles that will need to be resourced, each with a specific human to bot resourcing ratio:
- Dispatcher - 24/7 bot monitoring is necessary to ensure bots have run successfully as dictated by their predetermined schedules, or as needed in the case of “on-demand” or “invoked” bots. The monitoring itself is typically done as a dashboard component of whichever software platform your organization has chosen to build and deploy bots with. (Leaders in the space like UI path, Blue Prism, and Automation Anywhere all have their own monitoring tools as part of their application suite.) This dashboard will essentially tell you whether a bot has run successfully or not. If the bot fails to run successfully, the dashboard should be configured to push an alert, or generate a support ticket. The resourcing consideration here is, how many people do we need to manage these alerts / tickets, communications to the impacted parties, analysis of the root cause, and escalation where necessary? (Standard operating procedures for each of the aforementioned failure-response functions should be developed, but we’ll discuss that in a separate post.) The Dispatcher should be trained on procedures for triaging failures and the proper communication or steps to address and / or escalate. Development experience in this role is plus, but not necessary. Our recommendation is that one full time monitoring support resource, or Dispatcher, should be allocated for every thirty bots in production.
- Responder - Responders are the bot support personnel receiving an escalation from the Dispatcher who has been alerted to a failure that they are unable to resolve by following predetermined triage protocol, such as restarting the automation, or rebooting the VM it runs on. Responders should be software engineers able to reference documentation related to a specific bot, review the underlying code, and pinpoint the failure’s root cause. Once they have identified the root cause, standard IT “break / fix” protocols should be enacted. For the majority of companies we work with this process includes solutioning, development, testing, and deployment. Once a solution has been deployed successfully the Responder will notify the Dispatcher, who will update relevant parties, and mark the associated ticket “resolved”. We recommend at least one full time Responder for every fifteen bots in production.
- Subject Matter Expert (SME) - Each automated process, or bot, will need a designated business SME who knows the previously manual process, and can advise on the intended bot functionality. Separately, and more relevant to supplemental support resourcing, an SME must be identified for each application or system that is touched by an automation. This will often be the “application owner” for your organization. The SME will help guide solutioning related to the specific system they oversee. For example, a Responder assigned to a bot failure ticket may uncover a root cause related to a recent update to the ERP system that the bot interacts with as a part of its intended purpose. The Responder will then need to bring the SME into the solutioning process to provide the breadth of knowledge of the ERP system necessary to resolve the failure. The SME does not need to be a dedicated resource, but should be aware of their role in the support process, and be informed of any new bots, or changes to existing bots, running on the application or system they are responsible for. This is a two way street, as the SME should inform Dispatchers of upcoming application / system maintenance or upgrades that could impact bot success, allowing Dispatchers to plan proactively and avoid production failures.
To summarize, role resourcing can follow these guidelines:
- 1 full time Dispatcher (monitoring resource) / 30 bots
- 1 full time Responder (break / fix resource) / 15 bots
- 1 part time SME (specialized resource) / each application or system touched
The above recommendations are high level guidelines that should provide a framework for resourcing, but should be adjusted as needed based on the criticality of the automated process, and complexity of the automation itself. Resources can be on or offshore, but should have pre-determined and agreed upon response / uptime service level agreements (SLAs) for each bot. It should be noted that many organizations start out with a federated support model in which each business unit manages the failures / exceptions of the bots they deploy. This may work initially with less than 30 bots in production, but it will quickly become a full time job for a centralized support team.
Have some feedback about these recommendations, or want to get in touch with us to see how your own RPA process stacks up, or for help designing SOP and resourcing / response protocols for your automation program. Email us at firstname.lastname@example.org