Published on the 16/08/2017 | Written by Donovan Jackson

From banking to botrytis, new ways of doing business better are an ongoing experiment…

Analytics, AI, data science. These abstract concepts fundamentally underpin the ability for businesses to serve multiple customers at low cost while delivering experiences which amaze and make individuals feel special. They also underpin one of the newer entrants to customer engagement, chatbots.

Ahead of the 2017 NZ Analytics and Insights Forum, we caught up with some of the country’s leading practitioners in what is a fascinating area of the tech industry. ANZ Head of Customer Insights Ratneesh Suri told iStart that applying analytics to drive service delivery depends on putting the customer ‘at the heart of everything you are doing’.

“Particularly for large organisations, different parts of the business tend to see customers through a particular lens which offers a partial view.  Customers, however, are looking for a seamless experience and want to deal with a single organisation. Analytics provides the ability to join the dots and create a 360-degree view using information from multiple systems, while providing insights which the organisation can use to build great experiences for individual customers,” she said.

While organisations like ANZ have millions of customers, each individual person is unique. We all like to feel special and looked after; however, if we all really were looked after as individuals, the cost of doing so in the past would have made services prohibitively expensive. “That’s really the power of automation and analytics in today’s environment; it allows for the delivery of personalised service to multiple customers,” said Suri.

It also provides the ability to seek out or create a competitive advantage, or to add further value to customers. At Wine-Searcher.com, AI & Analytics team leader Vikash Kumar is creating a chatbot to do just that, and said there are now ‘two Martins’ at the company. “There’s Martin [Brown], the founder and CEO, and there’s Martin the chatbot. You can ask either one of them for a wine recommendation, but the chatbot will soon be available to everyone, all the time.”

Martin the chatbot, said Kumar, will make recommendations based on a wide range of criteria: price preference, grape preference, region; “There are quite a few parameters you can put in there,” he confirmed.

But making Martin work isn’t as simple as plugging in a quick service. Kumar said the field is ‘very new’ and getting people to engage can be a challenge. “On an ecommerce site, people are used to buttons, not interacting with a chatbot; user testing has shown that while most people can work well with the chatbot, many were not able to find some of the useful features we have put in there. We are now redesigning those features to stand out better.”

This is among the simpler lessons which become clearer as chatbots are translated from theory into practice.

It quickly gets more complicated: “With a chatbot, people want to have a conversation rather than having a ‘search’ function – which is pure ‘ask and answer’ – so there has to be a memory of what came before to guide that conversation, and context to retain the thread of the discussion,” Kumar explained.

Gauging intent, something humans do quite well, is far more difficult for a machine to do, he added. And while memory is one thing, it is also necessary to forget some things for a conversation to move forward; people naturally stumble, make mistakes and test and discard alternate ideas through conversation. All that makes the creation of a chatbot quite challenging.

Or, as Kumar puts it: “Initially we introduced something we thought was simple, but people didn’t get it. But by working with Martin, things are developing and maturing, and while it is an exciting field to be working in, there are many journeys back to the drawing board to align with users’ expectations and experiences.”

That said, he added that there is ‘readiness to engage’. “Our testing with Martin shows that people are looking forward to it. If they get some results, they are not turned off; 8 out of 10 say they definitely will use the chatbot again.”

Suri said that the pace of development in the field of analytics is rapid, which makes the practical applications a constantly moving target. “Customers expect a whole lot more from the business now. If you aren’t meeting those expectations – and analytics absolutely plays a fundamental role in that – you’re not going to be competitive.”

She stressed the necessity for utility in analytics and its ultimate outcomes. “If your insights are not being operationalised and put into action, they are meaningless. Data is the biggest strategic asset for any business and analytics is the key to unlocking value from it.”

The 2017 NZ Analytics and Insights Forum takes place at the Grand Millennium Hotel on 30 August.