When Mark Zuckerberg unleashed his 2016’s yearly challenge of building a simply AI bot called Jarvis in Facebook, the whole world gaped in fascination that their sci-fi fantasies are finally here. Inspired by the movie Iron Man, the Internet Demigod had built a bot from scratch using a combination of AI, Machine Learning (ML) and Natural Language Processing (NLP) techniques to control several aspects of his home-from simple “grey t-shirt cannon”, temperature and volume control to advanced functionalities like allowing/barring entry to visitors by advanced image recognition, an interactive mandarin teaching program for his daughter etc.
The most interesting aspect about Zuck’s ambitious project is his plan to build similar AI products for the world to use thereby accelerating the path to Machine and Deep Learning applications everywhere in the world. From a business point of view, the most exciting use case would be applying ML to understand customers better than ever before.
Breaking barriers with Machine Learning
According to a Gartner report, “By 2020, customers will manage 85% of their relationship with the enterprises without interacting with a human”. Today, the path is clear for the next gen Customer Experience Management that would predominantly revolve around ML findings of the decade.
In the pursuit, one of the biggest obstacles is that the concept of Machine Learning, by nature, wasn’t developed to be applied straight into public-facing requirements. The clear demarcation around AI and ML also lies in the fact that the former involves feeding information and semantics from outside for systems to enable machines to make decisions on their own and the latter is taught to grow a brain that can feed and process data all by itself. So, when it comes to taking ML to the masses, there was a clear requirement to break it down to applications that can be operated by scientists and business decision-makers alike.
A 2016 Harvard Business Review report talks about the barriers that simple software developers had in breaking their way into the ML space and how natural progression has led to commercialization in this space, drawing parallels to every software development that happened in the last 50 years. Now, opportunities through “commercially available ML frameworks”- like Tensorflow, scikit in Python etc. – have been rolled into the market with the aim to let anyone deep learn their way around their data.
The Future of CX with Machine Learning
With exploding competition in almost every market, enterprises are hell-bent in nourishing their customers, right from the discovery phase to the purchase and even beyond, by feeding Machine Learning techniques to their systems thereby enhancing the customers’ overall experience during their association with the brand.
Some of the key areas where ML can make huge progress in better CX are:
- Increased Accuracy: Growing competition in the market also means that a brand has to attract a customer within the first few minutes of a casual glance. With ML, companies are looking for improving their aim at hitting the bull’s eye in terms of what that particular customer might expect out of their offerings. For example, American retailer The North Face has developed an ML interface called “Expert Personal Shopper”. This interactive platform registers the shopper’s preference with random phrases like “New York in winter” and “Women” to display not just an array of winter accessory choices but also weather predictions per month.
- Improved Personalization: In a research conducted by VoC Research, some of the top comments on personalization today by the consumers were, “They (businesses) aren’t personalizing things that matter to me” and “I want more than just buying history-based emails”. Modern customers are aware of the data that is being collected from them and as a result, demand a highly personalized, sublime level of engagement addressing their multi-dimensional requirements. Companies like Episerver are already making it happen through more than 100s of machine learning algorithms combined in a module called “autonomous personalization” targeting unique, personalized shopping to every single customer.
- Enhanced Customer Desk Applications: Another area where ML-based systems are expected to create wonders is the customer desk applications. While the deep learning systems wouldn’t completely wipe away the human representation, it can come incredibly handy for the customer service representatives through a multitude of features. For example with Machine Learning, CSR systems would be able to tag data with images, group tickets under tags, make advanced prioritizations and even resolve issues of limited magnitude instantaneously.
- Advanced Risk & Fraud Detection: One of the best applications of deep learning programs could be fraud and anomaly detection systems. In sectors with very low appetite for risks like insurance, machine learning algorithms offer personalization to the insurers with specific pricing risk & loss estimation metrics. In anomaly & fraud detection sections, ML programs aid insurance companies by identifying fraudulent claims and patterns through phases of continuous learning & predictions.
The gen-next of Customer Experience will not just be measured by the avenues that a customer can contact to deal with their queries but to empower them with data so that they need not look further. With Big Data and bigger data-led business complexities, gaming machines to make the most out of the bytes and opening up new avenues in understanding intricate human cognition patterns is the ideal way for accelerated customer satisfaction indices.
Latest posts by Bhargavi Seshadri (see all)
- What are the top DevOps trends to watch out for in 2018? - January 18, 2018
- Hybrid Cloud Management with ServiceNow - November 3, 2017
- Accelerate Margins with JDA Retail Planning Solutions- EP and Allocation - November 1, 2017