“Machine Learning” and “Artificial Intelligence” are concepts often discussed, but rarely explained outside of highly technical publications. While these concepts are highly related, they can mean very different things depending on the application. Here at Tethr, we work with both concepts for the purpose of improving the customer experience and deriving insights from call, chat, and case data.
Today, we’ll define AI and ML, then talk about how to combine artificial intelligence and machine learning in order to improve the customer experience.
What is machine learning?
Machine Learning (ML) relates to the ability of a machine to “learn” or adapt to new data, and to find new mathematical relationships within that data. It is this adaptive aspect of ML that sets it apart from other methodologies.
Machine learning algorithms are designed to solve the unknowns of a problem that is often not well defined and to apply those solutions. Examples, in this case, would include object recognition and stock market prediction. In these cases, there may be no predefined rules to follow. Instead, the machine has only data and a goal to achieve. The system must seek to identify the mathematical relationships between the data to be processed and the desired output.
Machine learning using artificial neural networks
So, for example, a stream of video images may be fed to an “artificial neural network” or ANN, along with data indicating whether a given target object is present or not. Over time, as the data is fed to the ANN, its weight values are modified to capture information in the video images that uniquely identifies the target object. Errors between the machine’s predictions and the desired output are fed back into the machine to modify the underlying weight values so that the prediction is improved. Eventually, the ANN should converge on a previously unknown solution, that identifies the presence of the target object in new images, with the highest possible degree of accuracy.
What is artificial intelligence?
Artificial Intelligence (AI) is a much broader field of technology than ML, that deals with the ability of machines to solve complex problems.
Examples of AI that are not classified as ML include game logic, such as path planning and optimization. Chess programs, for example, generally do not “learn” the game of chess. Rather, the rules are programmed into them from the start. These algorithms seek to explore and optimize the sequence of possible moves and counter moves, maximizing the probability of victory. While such algorithms may be highly complex and capable of very sophisticated behaviors, each game and move within the game is handled as a new problem and is not remembered.
Combining artificial intelligence and machine learning
The greatest power of such algorithms is achieved when you combine artificial intelligence and machine learning. In the ANN case above, the machine not only learns the unknown relationships between complex inputs and outputs but is also able to actively explore those relationships to improve the desired outcome.
Examples of this approach can be drawn from the technologies developed here at Tethr. We have used ML to train the Tethr platform to predict customer survey results indicating the level of perceived effort in a customer interaction. Detected customer and agent behaviors as well as other conversational data are used as input to train the ML. The result? The Tethr Effort Index (TEI). A unique TEI score can then be calculated for every new interaction, predicting how the customer would likely have rated it.
Simply knowing how a customer feels about their interactions with a company isn’t nearly as important as understanding why the customer feels that way, and what the company can do to make it better. This is where the AI side of things comes into play. By exploring the ML model as it applies to a given interaction, it is possible to see the effects of hypothetical changes to the customer service agent’s behaviors and how they impact the TEI score.
In this way, the machine can identify the optimal agent behaviors needed to maximize the TEI score for any given interaction or group of interactions. Using this information it is then possible to coach the agent about what they can do to improve their TEI scores and thereby improve customer satisfaction. The result of combining ML and AI in this way is a prescriptive approach to improving customer satisfaction.
What happens when you combine artificial intelligence and machine learning?
These methods are applicable to a variety of key business outcomes, including sales and customer service efficiency, as well. It isn’t hard to imagine how the applied use of ML and AI can lead to significant improvements in business performance and to a company’s bottom line.
It is our goal, here at Tethr, to leverage our AI and ML technologies in a way that leads to direct and measurable business improvement for our customers.