Artificial Intelligence Machine Learning and eCommerce – improving all the time
In recent days, ‘artificial intelligence’ (AI) and ‘machine learning’ (ML) has been in the headlines after Facebook shut down an AI system after it invented its own language, the system having decided that English was an inefficient and illogical way of communicating. Mention AI to the man in the street and the first image that will often come to mind is a stereotypical scene from The Matrix or Terminator.
Thankfully reality is rather different and offers merchants and entrepreneurs the potential to benefit from consistent, continuing improvements in efficiency and operations.
So what is AI / ML?
In broad terms, AI is a branch of computer science attempting to build machines capable of intelligent behaviour, while ML is defined as “the science of getting computers to act without being explicitly programmed”. To give a simplified example, when using the Google search engine and you make a typo in a search term, you get the message: “Did you mean…?” This is due to one of Google’s ML algorithms that detects what searches you make, recognises what typos are typical, and ‘learns’ to correct it (or rather, to offer a correction).
Uses for e-Commerce
AI has uses across a full range of sectors, but for e-Commerce some of the potential areas include:
- Increased personalisation (e.g. for emails, related products, cross-sells);
- Chatbots for customer service;
- Prediction software (e.g. for stock depletion, sales and demand forecasting);
- Analytics (e.g. to determine irregularities, caveats, space for strategy improvement); and
- Working in combination with blockchain, e.g. for automatic re-supply.
The overall benefits of using AI/AIML within an e-Commerce framework relate to greater efficiency, cost-effectiveness, and an improved competitive edge. To take one example, that of chatbots, these have been deployed by many large companies to act as a first-line gate-keeper to deal with queries – in effect, they filter out all the simple queries that can be dealt with by the chatbot AI programme itself, meaning that only complex issues need to be escalated to a ‘real’ human agent. Survey results from the Aspect Consumer Experience show that 70% of respondents preferred to use chatbots to interact with companies for simple to moderation transactions. The advantages of this are immediate: 24 hours‑a‑day coverage for help desks or customer service centres is possible, improved response time can be achieved, and overall increased customer satisfaction.
To give another example: small companies and e-merchants do not always have access to the range of tools or expertise that larger companies have for predicting sales and demand flows. AI can assist this, by crunching data and reporting to produce meaningful predictions. This is one of the key aspects of AI/AIML – data and information. Clearly any system needs something to work with. Properly used, AI can draw data from a wide range of sources from within a company, a sector or an industry, and make sense of this within an actionable timeframe, allowing – for example – better stock flow management. However, this is not to say that AI is a panacea: any use of AI needs to fit in with (amongst other elements) a coherent strategy, an effective product, and good customer service. But when deployed in line with these is can act as a multiplier to give the creative merchant an edge against their competitors.
Increased personalisation is potentially one of the quickest wins for e-Commerce merchants. Tailoring publicity or promotional material to the individual customer (or potential customer) is one of the easiest ways to increase conversion rates. Automation of marketing is already widespread, with systems able to respond to particular trigger events or in accordance with set rules. This has already allowed marketing staff to focus on the creative side of the business. The next step – and what is beginning to emerge now, according to a recent item from Forbes – is the use of AI increasingly within the creative side of the marketing field. This links heavily into the data crunching aspects, combining sales forecasting and predictive analytics, but the end result could be the emergence of automated marketing campaigns that are flexible and adaptive to events and respond to changes in sales patterns.
The use of AI in combination with blockchains is one of the potentially more interesting aspects that could be explored by e-Commerce merchants. The use of blockchains to improve the security of online transactions, combined with AI that is on the lookout for irregular transaction patterns, could well emerge as one of the best ways to reduce fraud.
Already deployed by large companies in several areas, AI/AIML will have an increasing role in the future. The costs and economies of scale are swiftly enabling these advances to be open to e-Commerce and smaller merchants, offering potential edges and advantages that are worth exploring.