Eye on eCommerce: IMImobile on chatbots

One of the biggest developments in customer service in 2016 has been Facebook's launch of its chatbot-manned Messenger platform. The launch earlier this year, provides businesses with another channel to deliver customer service using automated technology, which is supported by humans who interact and teach the bots via machine learning.

The technology is not restricted to Facebook, and can be embedded in a retailer's app or website as a way to engage with customers deploying the same casual language customers use to communicate with friends and family via WhatsApp.

But is this technology just another channel to add to a retailer's portfolio of engagement tools, which already include email, SMS and call centres, to name a few?

"Chatbots have been on the radar for a while," Aseem Sadana, SVP of products and strategy at IMImobile, tells Essential Retail. "It's yet another point of presence and another interface in the digital landscape."

While Sadana hails the innovative technology, he says retailers have to remember it is ultimately just another channel to be incorporated into a business's digital strategy.

"It's a convenient point of interaction, and the way it engages consumers is very useful, but one has to be considering what consumers will be drawn to it for – it's not a replacement for an app or another channel."

Marketing or customer service?

He says the technology lends itself well to booking an Uber or ordering a regular takeaway, but if someone wants to buy a suit from an online fashion retailer, it would be far more cumbersome through chat. That said, if the customer leaves the website before completing the suit transaction, that is when chat bots could be used to engage with the customer.

But does this open up chat bots to the world of marketing, as opposed to service? Will we be bombarded by retailer's announcing the latest clothing range or a sale everytime we log into Facebook to check-in with our friends?

This is where retailers need to tread carefully, says Sadana. "P2P [SMS] messaging wasn't meant to be commercialised necessarily, it was for talking to friends," he explains. "So retailers need to bare in mind context and proximity and the immediate benefit."

One advantage of Facebook's chatbots in Messenger is the example of two friends talking about going for dinner and they invite the Facebook page of their local restaurant into the conversation to book a table. "That's very contextual," explains Sadana. "It's a natural use case and the bot can take over with the view of two people talking about dining and book a table for two. But beyond that one has to be very careful of pushing too much."

He says retailers need to embed bots in the right context and at the right time – an abandoned basket is contextual, but an unsolicited 2-4-1 offer, we're familiar seeing in our email inbox, may not. "You need to test and learn and it's important to be driven by the overall digital journey and which channels make sense for what purpose.

"Even if you build a chat bot, how it gets discovered is important to consider," adds Sadana. "If you are a brand, people might search for you on Facebook Messenger for example and the first interaction might be inbound. But for most companies, the approach would be to notify a customer in relation to a digital journey started elsewhere, for example order fulfilment comms or abandoned basket."

He says long-tail players could benefit from new reach. "A dentist or a curtain supplier, who would never make it with an app on your home screen, can certainly proactively establish presence on mobile through this route.

"The route here is chat – bots themselves are an automation of that interaction. In my view, companies can do a lot more with multichannel, context driven chat with good old agents and establish the use cases for automation before setting up bots."

Too much hype?

But Sadana is concerned there has been a lot of hype about chat bots changing how customers will buy things. "I don't think it will be that revolutionary," he admits. "In reality it will be a gradual process where companies learn to adapt their overall digital journey to include another channel."

He says the impact for retailers will be far more on the back-end and the service-orientated architecture to support the different channels. "User interfaces will continue to change, but what won't change is CRM, order fulfilment and things like the simple principle of a retailer emailing a customer four times in one week – if they must send a fifth, it should be something different."

So retailers need to apply the same best practices they do to other CRM channels. "The back-end is where retailers should worry far more than being carried away from the hype of bots," advises Sadana.

He also believes the artificial intelligence technology being associated with chat bots is too overrated. "We don't have complete AI driving these bots yet, and it's certainly not successful even in trials. Even Facebook Messenger is operated by a large set of desks of people answering your queries in a concierge-type fashion," he explains. "The, idea of training algorithms and machine learning, and optimising algorithms for a future of completely mechanised artificial intelligence is still far away."

But over the next 5-10 years, as customers become truly mobile-first in everything they do, the rich interactions provided by chat bots will be taken for granted.

"The fact that messaging itself has grown as it has with the overall trend of smartphones in P2P and eCommerce, shows how quickly things have moved," he concludes. "And things like the percentage of mobile bookings has been growing steeply, and will only continue in the future."

Sadana adds: "Messaging apps remain the leaders in mobile first, and that will only grow and become core to behaviour with mobile content. And if messaging apps are at the centre and reinforcing itself as the real estate means that the commerce and interactions through messaging will become even more important."

For more information, click below: