The history of the chatbot: Where it was and where it’s going
Chatbots are not new. The software has been around for decades, and now its application is beginning to impact the enterprise — both internally and outward facing
The history of the chatbot
The first chatbot was developed in 1966 at MIT — they called it ELIZA.
ELIZA, the mother of all chatbots, answered some very simple decision tree questions.
This first chatbot iteration has since evolved and developed, naturally. Through the 1980s and 1990s, the technology was deployed in automated telephone systems that used very simple decision trees, right through to MSN and AOL.
As we move further along the timeline, chatbot technology has exploded across social and business channels. Why are we so interested in them now? Why has so much attention been placed on the technology?
“It comes down to three main reasons,” says Howard Pull, strategic development director at MullenLowe Profero.
1. Messenger apps and voice-assistance are both inductor bots
“If you think of the likes of Facebook Messenger, WhatsApp, LINE, they have really opened up their APIs, so brands can now launch and create e-services that do everything from product recommendations, booking and service through to FAQs within that.”
2. A change in a ability
“There’s also been a step change in chatbots’ ability to sense; the ability for these bots to recognise images and recognise voice, which has allowed us to essentially search using imagery. If you’re Amazon, for your parts service, you can search via that, or if you’re EasyJet, you can provide an image and let the bots search from that.”
3. The rise of AI
“Finally, there’s been a step change in the learning, and that’s where the question of AI fits in. The ability to appear human and deal with more human conversations has undergone a massive step change in the last 24 months, in terms of the intelligence and the machine learning that sits behind the chatbot. It’s allowing brands and services to create an interface that feels human and interacts in a way that people expect to be spoken to and dealt with.”
At the moment, there’s a misconception around the abilities of bots. There’s a very high consumer expectation that a bot should be able to answer open-ended questions. However, this isn’t always the case and can be frustrating — as I’m sure many of you have experienced.
The quality of chatbots differs wildly, depending on how much the brand has invested in it (from a technology and operational perspective)
“I think the number one misconception surrounds the perceived role of a chatbot: to chat,” says Pull. “The actual goal is to solve a problem, to provide a service and give the customer that service in the quickest possible time.”
This isn’t an exercise in replacing humans; it’s about finding new places and new ways to interact with consumers. And, the better quality the bot, the more likely a brand is to attract and retain their audience.
It also not about creating something that is massively conversational. “People really want something that provides quick answers, connects them to a service, allows them to buy something, in the simplest way possible,” explains Pull.
The chatbot experience
Chatbots, as we’ve explored this month, are going to play a massive role in improving the customer experience.
Looking at the customer service experience, there’s a huge amount of frustration with most brands when you try to contact them — waiting times, being put on hold etcetera. For a service brand, investment in chatbot technology is absolutely critical.
Moving beyond this into a brand that has purpose and is looking to change things in its sector, chatbots are also going to be critical for increasing engagement and retention levels.
The internal chatbot
In the past 12-18 months there has been an explosion of chatbots in the enterprise, internally. In this scenario, it’s about treating an employees similar to a customer.
“A chatbot can be used in everything from room booking to IT,” says Pull. “In a large, mobile workforce, the tech can provide a simple, quick interaction for employees to engage with the business. That’s where you begin to see the big players, such as Microsoft and what they traditionally do with Office365: how can we put in that chat-based technology, open some of those services up and really reduce the need to call and give those employees a much simpler, more effective experience.”
At the moment, this type of service is emerging in brands that have big manufacturing or big workforces. But, it will begin filtering out into the mainstream, broader business world.
The future of chatbots
As the technology that sits behind chatbots advances, the applications and potential of the bots will change in a number of ways.
They’re are going to get a lot smarter so you’ll see brands invest a lot in new areas beyond service.
“If we look at where the main use cases are, customer service has been a traditional one, just because the money is there. So if you’re a brand like Amtrak, you can save $1m a year on customer service expenses, just by automating that,” explains Pull.
The focus of chatbots has been on customer service, purely because there is a much clearer financial return.
Chatbots will begin to solve new problems — and this surrounds the idea of choice.
“With companies such as Diageo, we’re looking at how we can pair a whisky to your personal tastes, and bring all that together through a chat interface — Diageo Whisky Matcher,” says Pull. “So, we’re going to see chatbots provide more advice; I think that’s a great example of where the technology is going to move towards. Today, we’re still in that FAQ simple customer service stage.”
Bots like Cleo, for example, are helping people manage their money and providing personal financial advice. The bot lives in Facebook messenger, but is moving to its own app soon. It offers pre-emptive nudges and gives straight forward answers to budget based questions.
In the future, the software bots will also be able to learn. And, as a result, move into a much broader range of activities. The technology will have the ability to recognise voice and within that, tone. Through that, it can recognise if a customer is happy or unhappy, and react accordingly. This will be a game changer from the customer-facing side.
Chatbots and AI are going to integrate across channels.
“One of the things we’re advising a lot of brands to look at is how can they take the smarts that they’ve built into a bot and replicate that across their channels,” says Pull. “It’s not just about creating something for Facebook messenger or WhatsApp as a bot, it’s about taking that smart, taking the work you’ve done on that AI, and think how can we apply that to the voice service. How can we then apply that to the website, how can we then apply that in store?”
“Over the next few years, we’re going to see AI and bots packaged up and put into a lot of different channels, beyond text and voice.”
On top of the examples mentioned above, Pull provided Information Age with some other bots to look out for:
• Lego gift bot called Ralph
• NHS for wellness
• Visabot for immigration services
• DoNotPay — a bot designed to simplify the parking ticket appeals process. In its first 21 months the bot took on 250,000 legal cases and won 160,000 appealing over $4m of parking tickets
This article was originally published on Information/Age