RPA Technical Insights, Part 3: Assisted or Unassisted Robotic Process Automation: How to choose the right delivery model for your project

David BrainBlog

This is part 3 of a 22 part blog series by the leading experts at Symphony Ventures. It addresses how to choose the right RPA tools for your business needs. Drawing from our global team’s extensive knowledge in automation consulting, implementation, and managed services across a range of diverse industries, we’ve drilled into the technical criteria to consider when selecting which RPA software best enables your company’s digital operation strategy.

Read part 2, Making a Complex Automation? You’ll Need Layered Design.

The vision for RPA is to distribute and manage work across machines as a ‘virtual workforce. Being able to deploy RPA at scale within your organization is key to achieving this. Essentially the RPA tools in the market offer two modes of deployment. Assisted automation, where the RPA software automates other applications running on the employee’s desktop, or unassisted, in which RPA is deployed on numerous machines to run without the need for the automation to be attended. This blog covers the benefits and limitations of each model and advises on what buyers should consider when selecting the right RPA tool for them.

First, let’s start with assisted automation. This method first came around as a breed of productivity tool for the call center. Here, agents would be able to trigger a series of automated steps across multiple applications using robots on their desktops, usually accessed through a sidebar provided as part of the software. This approach was effective in the reduction of average handling times, resulting in savings and improved customer experience. It also meant that long, complex processes were replaced with single mouse clicks, greatly reducing the time it takes to train an agent. The issue with many of these deployments soon came as desktops became inconsistent across an operation and some of the rudimentary methods of integrating with the applications started to fail. These inconsistencies could be different graphic cards, different resolutions or display settings. RPA software has come a long way since then but it is true that surface level automation is still required for certain tasks within certain applications. For that reason, there are still times when inconsistency of desktop image causes automations to fail.

Assisted Automation is not limited to call centers, but it is well suited to these tasks because it really should only be used when real time human-system interaction is required. For all other times, unassisted automation provides a more optimal solution.

As mentioned earlier, unassisted automation is where automations run on machines which are not attended. By this we mean that the automation does not need an employee to go up to the machine, log on, trigger the process to start, observe its performance then close the automation when it’s finished. These steps can be automated and facilitated through dashboards, which provide a control room through which to allocate work to machines, adjust priorities and queues and intervene with a specific robot’s performance (if required). For all intents and purposes, these machines do not need an employee to perform the automation. This opens up great possibilities: the robots can work 24×7 and they can work through queues of cases, only alerting an employee when something goes wrong. In essence, unassisted automation provides us with the level of automation we expect from a new, optimally designed system, but remember, the same outcome is being achieved while the RPA platform is automating all types of legacy, mainframe, win32 and web applications without the need for replacing any systems or building custom integrations into each of the application.

Unassisted Automation does however have a single and significant drawback: it needs structured digital information and clearly defined rules and this means it’s not right for every occasion. It is possible to use smart OCR tools or voice-to-text engines to convert this input but sometimes people are required. For these cases assisted automations may be your best bet. Alternatively splitting the process into human processes and automated processes can be the best option leveraging Unassisted Automation. An example of this approach is where an employee interacts with a client to complete an order form, then the robot could be triggered to work in unassisted automation mode to perform all the required checks, place the orders for the parts, schedule the installation, produce the invoice, etc.

So by now we have covered the benefits of each type of automation and when they should be used, we’ve explained the issues that inconsistency of desktops can have on some automations and why this can cause a problem for assisted, but why not unassisted? Well, the answer comes down to the high growth in hosting these robots in Virtual Machines (VMs). The adoption of VMs was a real breakthrough for RPA and is the reason that in recent years we have seen a resurgence of the technology. VMs provide consistent standard builds that are ideal for automation. This, combined with new integration techniques and improved control center functionality have really accelerated the adoption in the last few years of the techniques that were first being deployed over 20 years ago.

If you are trying to choose the right model for your organization, then we hope this article has helped. The reality is that as you expand your RPA deployments you will likely need to cater for both types of automation. A few tools in the market offer the capabilities to support both, whilst some specialize in just one or the other. Our impartial advice is to not necessarily rule out a vendor’s product just because it doesn’t support both models. Our work with multinationals in most sectors has taught us that more than one tool is often required within an organization. Our advice is that instead of first trying to find the right tool, focus instead on finding the right project. Find a project where the stakeholders are aligned, the benefits are clear and the impact is tangible. From there, design your operating model, your business case and choose the vendor that best supports both. We’ve helped many clients do just that through our Future of Work Assessments and we’ve found that, when this approach is followed, the impact of RPA is far greater and the take-up across the organization far more rapid.

Be sure not to miss part 4 of RPA Technical Insights, a look into the management potential that Process Administration brings to any automation.

Share this Post