The rise of chatbots
"Soooo bored!" exclaims my new match on Tinder. "Have you tried pulling the fire alarm?" I suggest. "Soooo bored!" she repeats without acknowledging my response. I suddenly realise that like millions of other poor saps around the globe, I have just matched with a chatbot. Time to revisit the arranged marriage option. Then again, with my luck I’ll probably get matched with a bot for that too.
What are chatbots and how do they work?
Chatbots are Artificial Intelligence (AI) software programs that are designed to simulate human conversation. In basic form, they take written or spoken text and match it with the program’s database in order to provide an intelligent response and engage in conversation or perform simple tasks. Essentially they are computer programs that you can chat with to achieve a result. Apple’s virtual assistant Siri is probably the best well-known example of a mainstream chatbot.
Traditionally, AI programs have had a lot of hype associated with them and expectations always exceeded delivered capability. As a result, mainstream AI was ignored for many years and only left to the technically competent ‘boffins’ to pursue.
And pursue it they did, through the ImageNet Challenge. The competition tests AI capability, where a computer is assessed on its ability to recognise and label images accordingly.
The closest a computer got in 2010 was 72%, compared with a human score of 95%. Then in 2012 a computer managed 85% by using an AI process called ‘Deep Learning’. Finally, in 2015 the accuracy went up to 96% when a computer beat a human. The results showed the world what AI was capable of and suddenly people started paying attention.
Why are chatbots important?
A recent study from BI Intelligence reveals that the use of messaging apps has surpassed the use of social networks. This means more people are having conversations rather than spending time broadcasting and consuming information on social networks. Remember, attention is the most coveted asset online, so those who invest in appearing where their customers already are will be miles ahead of those who try to attract them to new platforms like apps.
The Chinese application WeChat is a great example of this. Outside of China the value proposition is negligible, however within China, given its strict rules and regulations, WeChat is essential to daily existence. Locals in China use the app to shop, order transport through Didi Chuxing (the local equivalent of Uber) and even book travel. This is known as ‘conversational consumerism’ and presents an incredible opportunity for switched-on brands to leverage the concept.
Attention is the most coveted asset online, so those who invest in appearing where their customers already are will be miles ahead of those who try to attract them to new platforms like apps.
Digibank in India is another example of taking chatbots to the next level. The bank is staffed entirely by highly intelligent chatbots capable of allowing customers to perform banking services and responding to customer service inquiries. Digibank is powered by the KAI artificial intelligence platform made by Kasisto, the same company responsible for Siri.
What could go wrong with bots?
Earlier this year, Microsoft introduced ‘Tay’, a bot designed to engage with Twitter users to learn the secret language of millennials. Unfortunately the bot was exposed to undesirable influences and ended up learning nasty phrases that transformed it into an anti-feminist who actively supported Nazism. Microsoft soon shut down the project. It seems that, who your bot learns from is quite important in the grand scheme of things.
The other issue to think about is transparency. I hate thinking I’m talking to a human when it’s a bot instead.
If there was a notification upfront that the person on the other end is a bot, then that would make authenticity a lot easier. Remember, trust is still the number one currency of the new social economy that we live in and despite clever technology, you still have to respect people.
What is the future of chatbots?
The future is quite promising for chatbots, as smart companies will begin investing in communicating their value propositions through messaging apps. With advances in AI every day this means better quality conversations that create additional customer journeys.
There is plenty of information available online on how to set up your own bot, but successful implementation really comes down to 2 questions:
- Where can a conversation add real, tangible value to the customer experience?
- How can we train the bot to ensure it provides the best customer experience possible?
Answer these 2 questions correctly and you will win the hearts, minds and wallets of your customers. In this highly competitive environment, can you afford not to?