It’s good to chat: How bots evolved into many companies’ first line of engagement

If you’ve used any online service in the past couple of years, there’s a good chance you’ve interacted with a customer service bot.

Once an oddity, the automated assistants are becoming an increasingly common sight for consumer and businesses alike, serving up quick answers to relatively simple questions – a sort of interactive FAQ to help customers get the right answer quickly.

A bot can help you find extra discounts on your insurance, for example, or direct you to the right department online without a human ever having to be involved on the other side of the conversation. For the customer, that’s a win; for the business, it means they can keep customers happy without having to pile on the extra costs.

In the current economic climate, those automated assistants may be about to come into their own.

Sleep and rest are not the same and failure to address both can increase risk of burnout

---

Extend EV grant to second-hand purchases, says PwC

---

Card data suggests it’s not all doom and gloom in the hospitality sector

---

Intercom’s Fergal Reid has seen a lot happen in machine learning in the past few years. He joined the Irish-founded company five years ago, starting the machine learning team in Dublin. Since then, the company has built a lot of automation products, taking into account research and feedback from customers.

“We have some customers that use our automation to save a lot of money. With the economic climate, people are getting a little bit nervous,” said Reid. “We think that saving money on supports, on improving your customer engagement, retaining your customers is particularly important.”

There are good reasons why companies like the idea of chatbots. Unlike like customer agents, chatbots don’t need breaks, or even a living wage. And they don’t mind answering the same questions over and over again, freeing up the human customer agents for more complex queries.

Automation done right, done in a way that’s designed to help the user rather than wasting time, can be a win-win

“Automation done right, done in a way that’s designed to help the user rather than wasting time, can be a win-win,” said Reid. “You get your answer quicker, you don’t have to wait around and the person who’s running support operations can handle more cases, and their support team isn’t dealing with frustrated users.”

An increasing number of companies are recognising the value of their customers’ time. Google, for example, developed Hold For Me for its Pixel phones, allowing customers to turn the waiting over to Google Assistant. But a better option, you would imagine, is not creating the long queues to start with.

If you’ve ever had to wait on hold for a long time to have a simple query answered – and we all have, particularly in recent times – the idea of having a chatbot immediately available to weed out the easily solved issues from those that need the attention of a real-life agent seems like a sensible one.

[ AI helps chatbots to get better at chat, which helps us all ]

“I remember the first time we saw our resolution bot actually answer a person’s question,” said Reid. “Someone came up with a prototype we had live on Intercom, and asked a question. Before a support rep could even come into the conversation, the bot came and gave the perfect answer. That made me realise: ‘There’s something here, there’s a really interesting product here’.”

It’s not only businesses that can benefit. In January, NUI Galway blazed a trail when it teamed up with Chatspace to implement a bot as part of its students services options. That meant around the clock access for students to some services. Cara, the college says, is still learning; it is early days for the digital helper.

But chatbots have come a long way in recent years, as machine learning progresses. Who could forget Microsoft’s Tay, an experiment on Twitter that resulted in the tech giant pulling the plug on the Twitter-based chatbot after interactions with real-life users saw the bot spewing bigoted material. Microsoft was forced to shut the experiment down after 24 hours.

So chatbots with a specific purpose, designed to help companies and customers, rather than generic interactions, seem like a much safer bet. “There’s this sort of evolution. The machine learning, the AI is getting better and better, it’s getting real. You can start to build real products with it that you couldn’t dream of five or 10 years ago,” says Reid.

The first of these bots was designed to be simple and useful rather than annoying, Reid says, and would Answer questions with information. “For a lot of customers, maybe that could answer maybe 20 per cent of their support volume, or 40 per cent in the really great cases.”

[ Travel chatbots: What are they and do they really work? ]

From there, it was a case of figuring out what people did next to see if automation could help there. For some, it was taking an action, such as cancelling a booking, or amend it in some way.

“One of the things we built most recently is what we call custom actions, which essentially is like a no-code or a low-code way of our customer and it’s using the Intercom bot, to have the bot actually take action,” explained Reid.

“So where do we go from there? Well, the technology is getting better and better, and having bots that really understand more and more complex queries.”

The bots are getting more and more convincing and more and more intelligent with what they can do

Reid is predicting a revolution in the next three to five years. “You can see in the in the research area huge breakthroughs all the time,” he said. “The bots are getting more and more convincing and more and more intelligent with what they can do. And a company like Intercom has to stay close to that cutting edge and be able to get actual products that our customers can use, and use that to deliver more and more business value.”

According to research published this summer by Gartner, chatbots are set to become the primary customer service channel for around a quarter of organisations by 2027.

“Chatbots and virtual customer assistants (VCAs) have evolved over the past decade to become a critical technology component of a service organisation’s strategy,” said Uma Challa, Sr Director Analyst in the Gartner Customer Service & Support practice. “When designed correctly, chatbots can improve customer experience and drive positive customer emotion at a lower cost than live interactions.”

However, not all companies have the resources available to devote to creating these chatbots, either because they don’t have the expertise on hand, or their experts are busy elsewhere. And some of the companies that could most benefit from the technology are in the small and medium-sized business sector – Intercom’s one-time focus for its services.

Consequently, the simpler it is to implement these AI-powered bots, the more likely it is that companies will turn to them to help power their business. Intercom is adopting the low-code approach here.

“Even in big companies that have lots of engineers on staff, the particular person who is implementing intercom doesn’t necessarily have access to an engineering team,” Reid said.

“That’s how we ended up with that, building very low-code, integration custom actions or bots. I think there’s a real movement here. This technology always starts out difficult and hard. You need a great deal of specialisation to use it.

“Once upon a time, trying to use a car was this really huge, big thing. And we made it simpler and simpler. And as you make it simpler, more and more people can use it and you get value from it. And so I think the low-code movement you’re seeing, and technology generally, is a way of trying to open it up to a lot more people.”

[ Automated bot that calculates local VAT in less than five seconds ]

But with the development of the technology comes responsibility. Like Microsoft’s lessons learned from Tay, the increasing use of automation in their business must be carefully thought out.

“Speaking generally, this is an area that people are going to have to think more and more about. We are going to probably see bad bots, like spammy bots out there in the world,” Reid said. “Any sort of more advanced technology people can use it for bad things as well as good things. And so it’s, it’s important that people be thoughtful about this.”

There will also be questions over data security and privacy, largely governed by GDPR in Europe, as people become increasingly aware of the use of their private data and how it could be misused by bad actors.

But Reid, for the most part, is upbeat about the future of chatbots and automation in the Irish industry.

“I often think about the story of penicillin being discovered, but it was a decade later by the time it was actually productised and really saved lives,” Reid says. “So many things we do now in machine learning and AI, we’ve made the breakthrough but it’s probably going to be a decade of productisation. I think there’ll be a lot of exciting things.”