There are a few things that are almost universally true about IT service desks. First, most tickets go unanswered for at least a day. Second, when tickets are answered, only 70% are resolved during the first interaction. This means that of the hundreds or thousands of tickets your company creates each month, roughly 30% go more than a day without a resolution. While it’s likely that the most urgent tickets receive instant attention, the lag time for simple requests can have a dramatic impact on your organization. Employees can be locked out of computers or applications. Wi-Fi access could be down for multiple floors in your corporate headquarters. Your email could stop sending or receiving messages. Think of the cumulative strain even these common issues can have on your operation when left unresolved.
Even the most rudimentary Robotic Process Automation (RPA) tools can help solve some of these problems. By allowing automation to solve basic issues rather than relying on human labor, your business is reducing the time it takes for tickets to receive a first touch, and you’re freeing human workers to resolve more urgent problems. However, you will want to pick a tool that can have the most dramatic impact in as short a period of time and in a cost-effective manner.
In this article we’ll examine what to look for when choosing an RPA system, and why we think a solution like 1RPA, as part of our advanced autonomic platform 1Desk, is better equipped to help you achieve your goals.
Intelligent Virtual Agents (IVA)
If you choose a standard RPA tool that doesn’t come with an IVA, you will still be relying on human workers to triage IT requests. Humans don’t operate at machine speed, so requests get backlogged and crucial issues go unresolved. Yes, an IVA-less RPA tool can automate some common processes, but only if and when a human worker assigns it to a specific task.
By incorporating IVAs into RPA tools (as we’ve done with 1RPA), Artificial Intelligence (AI) becomes the first touchpoint for IT requests. An IVA can communicate with employees to help them resolve basic IT requests, and it can also triage and assign requests that require human attention. In the case of 1RPA, our industry-leading IVA Amelia automatically solves one-touch problems, such as resetting a password, requesting Wi-Fi access and more. This means on-demand resolution for low-level tasks, and on-demand triaging for more complex tasks. In other words, tickets no longer sit for a day or more before they reach the proper technician’s inbox.
Natural Language Understanding
What do the following characters have in common: 555 55 5555 and 555 555 5555? Both feature nine of the same numeric symbols. Both feature a space between the three sequences of numbers.
A bot studying these characters would be able to quickly discern that the first set of characters is a Social Security number and the second set of characters is a phone number. How? It’s because of the numerical sequence.
However, what happens if and when the numbers are entered in the following formats: 555555555 and 1(555)5555? This is where the difference between bots with and without Natural Language Understanding comes into play.
Simple RPA tools require data to be organized in highly structured formats. An email address must be entered into an email field, a phone number must be in a phone number field, etc. If data is in the wrong field, the bot enters it incorrectly into the process that it’s running. To fix this, manual labor is required to make sense of any data that doesn’t follow a strict field-based structure. In our example, if numbers are not placed into the correct Social Security and phone number fields, a simple bot will not be able to differentiate between the two. Even if a bot were smart enough to search for those two distinct styles of numbers, it would be thrown off by the lack of spaces or the parenthesis. The entire automation would require a human worker to structure the inputs.
The best bots such as those created by 1RPA can read web pages, emails, programs or any unstructured text put in front of them. They organize the data as required by a business process and deliver the correct placements for an automation. This happens on-the-fly with no need for you to limit your automations to structured formats. By leveraging a human-level of cognitive understanding of inputted information, both in structured and unstructured formats, higher-level bots can use their intelligence to ensure that each set of numbers goes in the correct spot.
Some RPA tools charge extra when volume spikes. If you have RPA resetting 1,000 passwords, most bots will need to reset each password individually until the task is completed. What if you need the passwords set at an even faster rate? You would need to deploy the same bot twice to cover 500 passwords apiece. Some vendors use this scenario as an opportunity to charge for the same bot twice. So even though robots are doing the work, you’re being charged as if you went ahead and hired extra workers.
The more advanced RPA systems offer bots that handle multiple tasks at one time. These systems are as scalable as the compute power you provide, so you can take a single bot and run it as many times as you need to in order to fulfill a request. Instead of running two bots for 500 password resets each, and paying twice as much as you would have for a single bot handling 1000 resets, you can run as many bots as needed simultaneously, and pay as if you only used one.
Be cautious and deliberate when selecting your RPA vendor. Find a solution that offers the most capable feature set available (which in our opinion includes cognitive features such as 1RPA). Most vendors offer an automation solution that can run simple tasks with lots of human oversight. Very few vendors can deliver the competence and sophistication of what we’ve described in this post.