Large-language models (LLMs) such as the one used by ChatGPT are now household names, impacting people across the globe in and out of the workplace environment. However, as ChatGPT’s data comes from scanning many sources on the internet, it can lack the specificity needed for tailored industry use cases. Industry-specific LLMs, such as those that draw their data from business conversations in customer services for example, will become more commonplace in the year ahead.

‘’An LLM is only as smart as the data it draws from. A specialised, domain-specific LLM will generate a more tailored and relevant output, while a more generic LLM has a higher likelihood of delivering hallucinations and biased results,’’ O’Connell said. ‘’Particularly in customer services, where real-time insights play a crucial role, this type of LLM will really become a game changer, empowering businesses to provide exceptional service and stay ahead in a rapidly evolving market.’’

Increasing diversity within organisations should always be a priority for business leaders, and the same can be said for data that is used to train AI. To ensure that bias is not built into AI models, leaders will need to ensure that they have diverse data teams that are bringing their distinct worldview and experience to the table.

‘‘Data is the fuel that powers AI, and if we don’t make a purposeful effort to ensure that training data is not biassed, we risk training our models to be as skewed as society can be,’’ O’Connell said. ‘‘This is something that needs to be recognised early on, which is why we have put guiding principles in place for our AI teams, as well as ensured that the teams themselves come from diverse backgrounds and schools of thought.’’

Individually crafted experiences for customers will be a defining trend of 2024. Companies will need to understand how AI can drive the most engaging online experiences for their customers or risk losing them. It’s well known that AI has the power to transform processes and make lives easier for workers, but we also need to understand how it can be used to take customer experiences to the next level.

‘‘Today’s customers expect an end-to-end digital experience, and they have high standards and expectations. Personalised recommendations and self-service options targeted at each customer’s unique needs are no longer optional - they are mandatory. If companies can keep up with these ever evolving market expectations and make use of new technologies, they will be able to keep up the pace and build customer loyalty,’’ said O’Connell.

Like people, AI models need training and development. They need to be constantly improved and perfected to make sure they are generating the most accurate outputs for a business, and as more companies adopt AI as part of their operations - this will need to be a key focus.

‘‘Supervised, active learning for models is a huge role that humans play in making sure AI systems are providing accuracy for end users,’’ O’Connell said. ‘‘In communications for example, companies have a responsibility to not only make sure bias does not come into play, but also continuously update the training data to account for new languages, dialects and forms of speech.’’

One process that can be incredibly time-consuming for organisations is onboarding new employees. During a period when there is considerable turnover in many industries, AI has the power to streamline onboarding in ways that would not have been possible a few years prior. 

‘‘In the era of generative AI, we have to re-think some of the traditional processes and figure out how to simplify them - onboarding is a prime example of this. Now, AI can generate information instantly for employees to support them through the process of joining a company. This reduces the resources and time needed to bring new employees up to speed with a company’s products and services. In the coming year, we’ll begin to see companies taking greater advantage of this and reaping the benefits,’’ O’Connell concluded.