Considering we started talking about AI almost two decades ago, it’s perhaps surprising that it is only just starting to make an impact on enterprises today. Certainly, this is the case with retail — today’s omnichannel shopping environment has placed a premium on efficient and relevant interactions with brands.
Retailers recognise that AI, and specifically machine learning, has the ability to handle vast amounts of data and is able to use that data to identify patterns and to make decisions with minimal human intervention. In today’s market conditions, this is an extremely appealing proposition; to be able to deliver more relevant shopping experiences whilst increasing operational efficiency at the same time. However, in many cases the anticipation of AI is still greater than its actual impact on day-to-day life for the vast majority of retailers.
There is a reason for this. While there’s no doubt that sound data practices are at the heart of AI’s success, too many fast-track the foundations of data science, and leap straight into AI, expecting to plug it in and for the algorithms to deliver out of the box. A black-box thrown into the e-commerce tech stack can certainly do a lot of automated heavy-lifting, but there is still a need for human intervention to guide what the algorithms are trying to achieve as well as augmenting their outputs with human ingenuity and inspiration. Delegating this responsibility to an opaque black-box to make all the decisions is short sighted – as the algorithm is only a part of the process. It cannot define what data to assess, how that data should be featured and the interpreting of the results in-line with commercial goals.
For Amazon, AI is a staple of their immense offering; perfunctory, efficient and relevant shopping interactions based on their AI-enabled ability to crunch through massive volumes of commercial, operational and shopper data. However, not all shoppers are seeking the commoditised Amazon experience for all their shopping needs. The challenge with the Amazon experience is that the data is not always as good as it should be, with search results delivering all kinds of irrelevant options whilst at the same time the shopper receives very little in terms of inspiration. This is a great example of why a broader, more expansive approach to smart automation needs to be taken and this is what sits at the heart of data science.
This doesn’t rule out AI’s value, but what is required is collaboration by design between human teams and machines. The human retail teams will need to identify their specific strengths and weaknesses to implement AI in a way that both engages with shoppers and makes commercial sense. In turn this will force greater cross-team collaboration in efforts to realise the true value of its investment.
Data Science in retail is the catalyst for this collaboration — it provides the foundation for an organisation to integrate its key business functions in a way that delivers the very best outcomes for the business as a whole. It all starts with asking the right questions. How are we seeking to differentiate our offering and deliver an exceptional shopping experience whilst gaining competitive advantage? This takes us to data; what data do I need to collect, blend and transform? And from this how do I curate and apply the insights from this data in a way that will help achieve the commercial goals of the organisation?
Retail has some of the most complex organisational structures among enterprises and many are not set up for a collaborative process as yet — after all siloed departments and teams don’t naturally work together, unless there is a clear reason to do so. When you place data and the corresponding insights at the centre of commercial decision-making, this approach is transformed. The retailers who are able to do this, and as a result deliver on AI, are the ones who are engaging with data science and scientists to erode the barrier between business functions. Where AI is able to automate, data science in retail creates a commercially-minded collaborative environment to discover and realise new opportunities.
Data science is becoming increasingly important to organisations trying to better harness AI, prized for its potential to unlock insight across teams that will ultimately influence whether a shopper wants to buy from a brand or not and remain loyal, and deliver return on investment — ROI. By helping to uncover trends, data scientists can also help organisations identify new drivers of key performance indicators (KPI’s) to make smart decisions, quickly.
Retailers who deploy AI or machine learning deployed in isolation of a data science-led environment is high-risk, at best. Without a common data strategy that allows all teams to think collaboratively, AI-led automation becomes a short-sighted effort. Shopper experience is something that is constantly evolving, and the definition of what is distinctive to consumers will by default change. By creating the foundation of a retail business built on data science practices, change doesn’t become a threat. With a joined-up strategy, retail teams will have the agility to evolve and collaborate with machines, and then AI will then become the promise technologists like myself have been waiting for.
Source: Information Age