Imagine a world where data from your fridge, cross-aligned with information about your spending behaviour, meant that just as you are about to run out of an everyday item, a delivery turned up to replace it. Imagine being permanently liberated from food shopping. Here, Lindsay McEwan, VP & MD EMEA, Tealium, tells RetailTechNews that while for some this is the stuff of futuristic dreams, for others it is a line retailers should not cross. As automation becomes more prevalent in customer service, mastering the art of data-driven and responsible convenience is increasingly challenging. Not least because there aren’t yet any defined rules for how artificial intelligence (AI) should be used.
Many industries, including retail, are therefore caught between eagerness to harness the benefits of AI and uncertainty over how to proceed in the right way. What companies need is an interim path that allows them to operate ethically while regulation catches up with technological progression.
Let’s explore how they can find one.
What does AI bring to the table?
To begin, it’s important to assess how AI can help retailers better serve their customers. The key advantage it offers is large-scale data analysis. As the number of connected devices rises, with 8.4 billion in use last year alone, so does the quantity of data they generate. Yet, while vast streams of information are a valuable resource for personalised services, filtering them quickly enough to meet expectations of real-time response and relevance can be difficult.
This is where AI comes in. Leveraging AI subsets – such as machine learning that can assess huge data stores and spot behavioural patterns – allows retailers to rapidly identify, and meet, unique needs across channels. Plus, AI enables retailers to give customers more choice about how they interact via instant communication tools that are available 24/7, such as chatbots. It comes as no surprise then that 61% of businesses said they implemented AI in 2017, up from 38% in 2016.
Create smart technology best practice
Despite the potential of AI applications to enhance efficiency, connectivity, and convenience, the lack of guidelines means there is room for unintended outcomes; see the Facebook bots that developed their own language. So retailers must set their own bar for what is and isn’t acceptable, starting with the most vital factor: data. AI algorithms are only as good as the data that fuels them – consequently, ensuring practices are effective and ethical requires access to clean, objective, and compliant data.
The measures retailers can adopt to achieve this are many and varied. Firstly, it’s advisable to guard against the errors fragmentation can cause by checking that their data is unified. By consolidating information held across different systems into one universal data hub, they can create a single source of truth for AI tools to work from.
Secondly, adhering to data regulation such as the GDPR is crucial. Designed to not only hand back power to individuals, but also establish clearer ground rules for responsible data usage in the digital age, such legislation is vital to create a safer future where relationships between brands and consumers are founded on openness and trust.
Finally, we come to reducing bias. Specifically, this is the bias that can come from humans, not technology. For instance, if a non-diverse workforce manages the data used to drive AI, there is a risk unconscious biases could influence output. As a result, retailers must eradicate any possible areas of bias and optimise workforce diversity so that datasets provide a heterogeneous foundation for impartial AI-aided decision making.
None of this, however, is to say progress towards better AI management hasn’t been made. The government, for example, has formed the AI Intelligence Committee to identify and address potential challenges, as well as the AI Sector Deal. Stipulations in the GDPR will also have an effect on how AI platforms handle data, although this will be limited to the obtainment, storage, and deployment of personally identifiable information (PII).
All of these initiatives are positive, but more work is needed. For retailers to perfect the art of convenience and use AI as an ethical means of improving customer experience, specific and comprehensive regulation is essential. An ‘automation’ line must be agreed that protects customers and puts their needs first by defining where the division lies between engaging data-powered communications and overly personal messaging.
Until then, retailers must set their own benchmark for what constitutes good AI usage and avoid overstepping the mark. By ensuring ethics are at the core of company-wide decision making, treating customer information with dignity, and exclusively utilising quality data, they can set a high standard that leads the way for the rest of the world.