Artificial Intelligence is not a silver bullet it is a tool. This tool however holds the key to fundamentally changing the way support is delivered. With the right application of AI fewer resources can handle more interactions and potentially in a more proactive and productive way.
What is AI
AI, or Artificial Intelligence is a convenient term that is widely accepted to represent access to a new type of intelligence – not artificial, but machine based. The promise of AI in in its ability represent human knowledge, make recommendations or predictions about something based on what it understands, and to extract meaning from vast amounts of data. AI can process and represent information with speed and accuracy at a scale beyond human capacity.
One of the pioneers of AI, Arthur Samuel, described it as a “field of study that gives computers the ability to learn without being explicitly programmed.” Samuel coined the term machine learning, which more accurately represents the state of AI today.
For the purposes of this article we will use the term AI to represent a new type of machine intelligence that we can leverage as a tool.
Types of AI
Today’s AI technologies are focused on representation of human knowledge typically within specific focus areas. Machines are taught a set of rules to perform actions or represent knowledge. Current AI is classified as Narrow AI. Strong AI is the next level of machine intelligence where rules, knowledge and learning enable machines to create new knowledge, solve unfamiliar problems and operate in ever-changing circumstances.
Examples of true machine intelligence exist in Hollywood and perhaps in the research labs, but for the purposes of applying it to post-sales services, we are limited to narrow AI capabilities.
Making Machines Intelligent
The reason we have introduce AI into the services lexicon is that the ability to make machines intelligent is now possible without a team of PhDs on staff (although some companies do have these types of resources on staff). AI-enable technologies provide services teams with the ability to introduce intelligence into the service process without having to develop customized software and algorithms.
Whether AI is enabled through custom software or out-of-the-box solutions, there are common elements that make AI possible.
For AI to act intelligent it must be taught about the domain in which it will be working. For technical support, a machine must be taught about products, operating environments, issues and solutions customers are likely to encounter. Machines may also be taught to identify circumstances that can cause a problem to occur. We may also choose to teach machines to understand customer sentiment so that we can monitor the health of customer relationships.
Teaching machines may involve defining a set of rules for it to operate against. Machine learning may also involve deep learning techniques where a large collection of data is provided as a basis for the machine to infer meaning from the words and structure of the data.
Whether through an out-of-the-box application interface or coded deep within the algorithms of a customer AI solution, teaching the machine is the first step to applying AI to support.
Natural Language Processing
The methods we use to teach machines to be intelligent will also require us to help machines understand our language. When machines can process and “understand” the way we speak it is referred to as Natural Language Processing. The natural language we speak and understand is more than just words, it is sentence structure and the meaning derived by how words are organized.
We must be certain that systems understand the nuance of our language and the lexicon of our domain. As an example, we need machines to understand the difference between how the word “crash” is used.
In a hospitality environment, crash may refer to a place to relax and may generally have a positive connotation.
It was a great place to crash for a night…
In a tech environment crash typically refers to an unplanned event with negative implications.
My computer crashed after installing the new upgrade.
For most AI systems, Natural Language Processing is a fundamental capability for receiving input and presenting knowledge. Fortunately, Natural Language Processing is commonly found in many out-of-the-box solutions.
Applying AI to Support
Teaching machines about our products and support-related issues and language processing (our support-specific lexicon) are foundations upon which we can build intelligent support applications. These basic building blocks make it possible to leverage machines for the delivery of technical support.
Machines can be taught to recognize patterns and concepts and respond with appropriate actions. In some cases, AI can identify and help prevent issues before they impact customers (e.g. analyzing telemetry from system and application monitoring). Some of the benefits of AI for Support may include:
- Machine intelligence can “understand” a user’s need even if it is not expressly stated.
- AI can identify customer sentiment expressed in cases records and customer feedback.
- Knowledge can be recalled by expressing a need or concept. No longer do we need to depend upon matching a search term or phrase within a collection of written human intelligence.
- Machine intelligence can interact with a user to collect necessary inputs to resolve issues.
- A machine can execute the most logical action in response to a user need such as presenting the right answer or routing a user to an expert that can help.
- AI can identify patterns that may cause problems and help to prevent issues.
- Machines can see the things that humans may miss (data analytics).
Making Support More Human
With the right application of AI fewer human resources can handle more interactions and potentially in a more proactive and productive way. While machines handle some of the support burden, support staff can focus on building and sustaining relationships through human-to-human activities:
- Develop plan to help customers adopt and apply products.
- Develop personal rapport with customers.
- Understand customer needs and expectations.
- Advocate for customer needs.
- Empathize when something goes wrong.
- Detect early signs of dissatisfaction.
- Maintain a personal relationship.
- Reinforce value of the relationship.