Kirsty Halloran, solutions manager at Avaya, looks at how National University Ireland Galway is trying to make artificial intelligence accessible for SMEs.
It goes without saying that the customer experience can make or break a business. If your agents are responding to a customer query and cannot find the right information quickly enough – or, even worse, give the wrong information – your customers can become frustrated, angry and defect to a competitor. This is why empowering your agents with the right technology and information is the key to building positive, intelligent customer interactions.
Artificial intelligence (AI) and machine learning (ML) play a vital role in enabling these interactions and enhancing agent knowledge. They can support human decision-making, simplify operations, and automate processes; all leading to a more intuitive and efficient customer experience. However, this technology often comes at a high price; limiting small to medium-sized enterprises (SMEs) ability to compete with larger, more well-resourced and funded enterprises.
“Artificial intelligence (AI) and machine learning (ML) play a vital role in enabling these interactions and enhancing agent knowledge”
Machine learning essentially happens in four steps: data is collected and curated; it is analysed, and models are built; predictions are made, and these predictions are then utilised and monitored.
The results are tested, and the model is fine-tuned through subsequent iterations to enable it to make more accurate predictions and analysis. Unfortunately, the collection and analysis of data is expensive and time-consuming.
Contact centres generate vast amounts of customer data across many touchpoints, social channels and various systems that contribute to the overall customer journey. While we all know what successful outcomes look like – for example, a satisfied customer or a purchased product – and what an unsuccessful outcome looks like.
What many companies struggle to understand is the more nuanced elements of the customer journey that can to lead certain outcomes. Or, alternatively, if there are timely interventions that they could take on that journey to make a successful outcome more likely.
Accurate predictions of outcomes are therefore the first step to planning interventions on the customer journey. Contact centre data points available are vast, but usually well labelled and structured, which should make data curation and algorithm training a little easier. However, data varies considerably by verticals such as retail, hospitality, or finance. The resources required to undertake AI training are often out of reach to small and mid-size organisations.
As a result, data analytics is a massive global research effort aimed at taking the guesswork out of decision making in society. The Insight SFI Research Centre for Data Analytics at National University of Ireland Galway, for instance, conducts research that enables the use of massive amounts of data to enable better decision making. In this environment, businesses can work with academia to come up with solutions to problems they encounter.
The research being undertaken by Insight and Avaya Ireland, showcased at the ITAG AtlanTec festival in summer 2020, discussed methods in which the curation of datasets, feature discovery, algorithm selection, and tuning can be automated – thereby enabling the business benefits and reducing the expertise required to roll-out.
“AI and machine learning are increasingly at the heart of contact centre solutions experience across all channels of communications, enhancing customer experience and driving operational efficiency”
There are different approaches that can be taken to machine learning including the deployment of Automated Machine Learning (AutoML) technologies from companies such as Google, TPOT or H20, but these can be burdensome for SMEs to deploy and manage. They still require specialist skills that may not be available outside larger enterprises.
The Avaya-Insight research focuses on the deployment of genetic algorithms as the strategy for machine learning – using a survival-of-the-fittest, evolution-inspired approach to understand a complex dataset. Rather than relying on humans to set up the machine learning environment, this research starts at a random position and allows the machine to find the structure itself by evaluating how well algorithms perform in terms of their ability to predict outcomes.
Selection is then applied to keep the most effective. Crossover and mutations are applied, and the next generation of algorithms are born and evaluated. This happens through iterative generations. Initial research on small samples has led to up to 20 per cent improvement in just five generations. This is a different approach to classical machine learning.
ML in customer experience environments can potentially predict call outcomes and automate processes or better route calls to agents for optimal results. It can learn customer trends, sentiment and areas for improved quality monitoring as well as identifying anomalies and targeting fraud. It can also be used to identify a company’s best agent behaviours to reward and identify new and upsell opportunities.
“AI and machine learning are increasingly at the heart of contact centre solutions experience across all channels of communications, enhancing customer experience and driving operational efficiency. Our goal is to ensure customers of all sizes can leverage this paradigm shift. We are partnering with Insight, one of the largest dedicated AI and Analytics Research Centres in EMEA, to develop innovative ways of bringing these technologies to as wide a customer base as possible,” according to Mike Conroy, vice president, Avaya Solutions and Technology.
With continued and dedicated research into AI and ML and partnerships between businesses and academia, businesses both large and small can discover relationships in past datasets and apply optimal learning algorithms in an automated way to make the benefits of AI accessible to organisations of any size.
By Kirsty Halloran, solutions manager at Avaya
Published: 17 November, 2020