Research from EMA shows 7% of IT people don’t make a distinction between Virtual Agents and AI bots—but there are differences. To make quick progress with AI in ITSM, you need to make sure everyone is on the same page.
Part of our Understanding AITSM series.
The most common use case for a virtual agent is to replace a human service desk agent. A virtual agent interacts directly with service customers. AI bots are the back-end workhorses that can automate much of the day-to-day work that IT people do—particularly IT operations and service delivery teams.
Virtual agents do what human service desk agents do: communicate with IT customers, work out what the customer’s problem is, and either take action to solve it or help the customer take action. To understand the customer’s problem, a virtual agent must be able to take input from the customer (by text or speech, but most commonly by text), and then use Natural Language Processing (NLP) to make sense of it and decide on a response.
Confusingly, most virtual agents are presented to customers as a “chatbot”—simply because that is what virtual agents are usually called in the consumer world. Companies like Starbucks and Dominos have customer-facing chatbots on their websites. Some organizations have embedded chatbots into their mobile apps. To many people, the availability of a chatbot as a service/support channel is now an expectation and a preference. We call it a virtual agent chatbot, because we’ve found that whether people call it a virtual agent or a chatbot, everyone understands virtual agent chatbot.
It could be argued that a virtual agent is a type of AI bot. You could also say that virtual agents/chatbots are customer-facing AI bots, with an emphasis on human interaction. A virtual agent is machine-to-human (M2H), an AI bot is machine-to-machine (M2M).
A virtual agent is an example of AITSM—the application of AI to make IT service management better. We will be discussing AITSM in our understanding AITSM series over the next few weeks, so make sure you subscribe to this blog to get notifications.
In the context of IT, AI bots are the application of artificial intelligence to do the work that IT people do behind the scenes. They feed on data, analyse it, and act on it—either by alerting the right people, or by taking direct and immediate action. Their ability to analyse the mountains and streams of data goes far beyond human capabilities—meaning they can find patterns and insights that people could never find. The pools of data are too big. The streams of data are too fast. The complexity of IT management is now too high—it’s a big data problem now, so new architectures and technologies (like AI) are required to handle it.
AI puts real-time operational analytics within reach of most organizations. It also enables “zero touch”, detect-and-correct automations that work independently from people to reduce IT workloads and improve infrastructure resilience. Typically, specialized AI bot technologies will be deployed to cover different types of work: such as break-fix, IT agent augmentation, demand analysis, and knowledge curation.
For example, a bot can monitor infrastructure status data, use Machine Learning (ML) to understand how certain infrastructure conditions influence system downtime, predict these failures before they happen, and act to resolve issues before they impact services.
Typically, AI bots mine structured data to find patterns. However, some AI bots may (like virtual agents) have some Natural Language Processing (NLP) capability, so that it can make sense of unstructured data.
For example, our Knowledge Candidate Automation bot mines unstructured data (such as free-text descriptions within incident records) to make connections between issues and solutions. These suggestions are flagged for human review before automated publishing to the knowledge base—driving growth of the knowledge base while reducing the amount of time IT people spend on knowledge management.
Our InfoZone agent assistant is an AI-driven bot that helps service desk agents solve issues. As the agent types, the InfoZone AI bot does real-time mining of ITSM and ITOM data in the assyst CMDB to make suggestions to the agent: the right questions to ask, the steps required to diagnose the issue, similar incidents, matching knowledge articles, and any other outstanding issues that can be cleared up on the same call. This greatly improves agent performance and reduces call volumes and times—a double win situation.
Knowledge Candidate Automation is an AI bot. InfoZone is an AI bot. Detect-and-correct is an AI bot. All these smart IT helpers are bots. By contrast, the AI that helps IT’s customers with their issues and requests is a virtual agent/chatbot.
KEY POINT: Virtual agents interact with people. AI bots work behind the scenes, interacting with machines to get things done.
Now you can distinguish between a virtual agent and a chatbot—and understand some of the nuanced terminology. But you can’t always guarantee everyone else in your organization will stick to the same rules.
We find that the context in which people use the phrases “virtual agent” and “AI bot” helps to pinpoint what a customer really means. By focusing-in on the problems they need to solve, and the specific use cases they want to tackle, it’s easy to understand how they are applying terminology and clarify definitions so that everyone is on the same page. Make sure you get the same clarity in your organization—so that everyone is pulling in the same direction.
Contact us for a no-obligation chat with a consultant
Sign-up for blog notifications to make sure you don’t miss out
More About Virtual Agents and AI Bots
- 6 Ways Automation and AI are Transforming Service Desks
- De-stress Your Service Desk with Detect-and-Correct Automation
- How Chatbots are Changing Life on the Service Desk
- How Advanced Service Desk Chatbots Work
- How to Launch a Service Desk Chatbot
- Service Desk Chatbots Rely on Quality Support Knowledge
- Service Desk Virtual Agent
- AITSM – assyst Intelligence for ITSM
- 64% of Service Desks Struggle to Find and Keep Good Agents
- AI in ITSM: 5 Angles to Consider