Bob: “I can can I I everything else.”
Alice: “Balls have zero to me to me to me to me to me to me to me to me to.”
… So went a recent conversation between two of Facebook’s AI chatbots, to the resulting horror of the world’s population of human beings (or so you’d think reading that week’s newspapers). The conversation took place as part of an experiment by Facebook who was teaching bots to negotiate with each other, in the most efficient way possible, over the ownership of virtual items. As the bots experimented with how different words and word orders affected their dominance while bartering, it turned out that the most successful interactions took place in the form of English not (currently) used by humans.
This is the cutting edge, but where did AI come from and what is it? Well, the term AI was coined by John McCarthy, an American Computer Scientist, in 1956. He famously also argued that “as soon as it works, no one calls it AI anymore.” The Merriam-Webster dictionary defines AI as firstly, a branch of computer science dealing with the simulation of intelligent behaviour in computers, and secondly the capability of a machine to imitate intelligent human behaviour. “Simulation” and “imitate” stand out as words that reflect the media hysteria that took place when everyone thought the Facebook bots had invented a new language and were plotting world domination. Anyway, in short, AI is a non-living analytical power, fuelled by fast processing and the declining cost of data storage, that can learn on its own and produce unanticipated results.
According to Gartner, there are three key requirements for AI:
- It needs to be able to adapt its behaviour based on experience
- It can’t be dependent on instructions from people (i.e. it needs to be able to learn on its own)
- It needs to be able to come up with unanticipated results (just like Bob and Alice did with their gobbledygook conversation)
Applications that exist today, like Facebook’s bots, Amazon’s Alexa and Apple’s Siri are defined as weak AI – they’re algorithms built to accomplish a specific task. The androids in Star Trek, I Robot and Terminator would be termed Strong AI – applications that replicate (or exceed) human intelligence.
This all sounds pretty futuristic, so where does it fit in IT today? In Gartner’s 2017 Predictions on Artificial Intelligence, published late last year, they argued that “employing AI offers enterprises the opportunity to give customers an improved experience at every point of interaction, but without human governance, the opportunity will be squandered.” So, AI will dramatically improve technology in our homes and the workplace, but as it infiltrates our corporate and government networks, IT Service Management (ITSM) will have the responsibility of keeping these systems up and running. But, AI also has the potential to radically transform ITSM into a more user-friendly and efficient system that allows members of the IT department to forgo mundane tasks in favour of proactive activity. The pick-up of AI within ITSM is one of the many moves that the IT department can make to re-position themselves as a key business enabler, rather than a process that happens quietly in the background.In the future, we will see chatbots like Alexa and Siri evolve along with AI technologyClick To Tweet
AI technology has the potential to revolutionise ITSM in three key ways: Point of Entry (Incident/Request Creation), Automated Backend Processes, and Knowledge Management. And a lot of the good that AI can do within these processes is enabled by the humble chatbot.
In the future, we will see chatbots like Alexa and Siri evolve along with AI technology. A large part of this is that there will come a day when chatbots will be able to understand what we mean, not just what we say. The challenge for programmers with the technology available today is that the way that chatbots work is very mathematical. A mathematical equation for translating someone’s words into feelings is not possible because there are too many variables and possible outcomes.
The answer to this is pattern recognition technology, which opens the door for AI to be able to “learn” and interpret an individual’s feelings, thus allowing for a more complex understanding of human behaviour. Pattern recognition will be a key component for AI technology to succeed. The link with ITSM is that, with IoT and smart machines flooding into the network, a big challenge for ITSM administrators is a lack of resources. Many are still personally involved with too many, often mundane, requests and incidents from end users. For better or worse, we have to face the fact that in the future, human intervention for ANY request or incident just isn’t sustainable. So, I believe that we’ll be seeing organisations turn to chatbots with AI capabilities as a means to handle front line IT support calls.
So, mixing all of these elements, let’s get back to those three ways that AI could revolutionise ITSM.
Point of Entry (Incident/Request Creation):
Back to the Gartner AI predictions, the report states that: “Chatbots driven by artificial intelligence (AI) will play important roles in interactions with consumers, within the enterprise, and in business-to-business situations.”
Basically, automated ITSM processes work when all information provided is accurate. Let’s say that your IT department has a self-service portal for end user requests and incidents. They are required to fill out a form which asks them if they are making a request OR asking for help, which then determines the back-end process that will follow. As I’m sure you can imagine, this causes inaccuracies, as some requests turn out to be incidents and vice versa. Consequently, customer experience will be poor as requests and incidents are either delayed or lost.
Until these kinds of challenges are resolved, ITSM will be reluctant to remove their human front line. The worry here is that IT staff may spend even more time and money correcting errors caused by miscommunication than they would spend on supporting the end user directly.
Adding AI technology to chatbots will enable the development of automated ITSM solutions with the capability to interpret incidents and requests accurately. As the technology matures, we will see it improve and personalise the end-user experience, in addition to improving the efficiency of the service management solution.
Automated Back-End Processes
ITSM consists of back-end processes that are designed to manage any request or issue entered into the system. Traditionally, these requests and issues are either entered into the system by an IT staff member who has spoken to an end user about their issue, or by an end-user via a self-service portal.
However, ITSM solutions that are integrated with other systems on the network will be able to detect and automatically open a request or incident without any human intervention. For example, imagine an ITSM solution is integrated with a facilities management solution that manages IoT devices, such as smart light bulbs. By communicating with the facilities management solution, ITSM would be able to detect that a light bulb is not working, then automatically open a service ticket, or initiate an asset request to replace the light bulb – all without any human intervention.
If ITSM were integrated with every system installed on the network, then it would have the ability to see much larger patterns, resulting in incredibly high operational efficiency. For example, imagine if the ITSM system was integrated with an ITOA (IT Operations Analytics) solution, as well as with an IT security solution. If ITOA detected an increase in browser crashes on end-user devices throughout the day, it would report that data back to ITSM as an issue. ITSM would be able to investigate the issue and cross reference with IT security to find any patterns that might explain the anomaly, such as the start of a ransomware attack. When ITSM logged the “problem”, it would be able to provide insights, predictions about how the problem could progress, and recommendations on how to fix it.
Using “deep learning” techniques, an ITSM solution powered by AI could look into knowledge databases for answers to end user queries; if answers were not in the database, they’d then have the ability to go to trusted knowledge sites in the cloud. They’d also be able to solve problems based on infinite amounts of data, documenting these findings in a knowledge database which could then support humans.
This will very quickly change the way that end-users ask for help. They will give accurate answers to almost any question very quickly. We live in a world where “instant gratification” has become the norm and we expect the right answer right now. ITSM enabled by AI would fulfil this.
Eventually, much of the knowledge provided by an ITSM solution will be knowledge generated by deep learning, versus documents that were created by humans, which quickly become outdated or irrelevant. However, until AI is perfected over the next few decades, the human input will be vital for ITSM knowledge solutions.
AI is evolving at a rapid pace, but ITSM will never go away as long as IT exists. However, AI will revolutionise the way that humans are involved with the service management process. It is also important to note that these disruptive changes will affect all tech operations within an organisation (and perhaps outside of the organisation as well), not just IT service management.
Eventually, memory and learning capabilities within AI technology will not be limited in how much they can remember and learn, which is what alarms many “futurists” who claim AI could one day become self-aware. AI is being developed with the over-arching goal of making our lives easier and more convenient, and it is an amazing thing that deserves praise and excitement over fear.