Just a few short years ago, customer service chatbots were stilted affairs, little better than the scripted agents they were supposed to replace. As AI has grown in stature and usefulness, those chatbots are now serving millions of customers for brands big and small. And the agents are still here, focusing on complex issues or those that require a human touch, and as 2020 beckons chatbots will be ubiquitous.

Back in 2017, the likes of IBM were quoting estimates that Chatbots will save businesses $8 billion each year by 2020. And it looks like that number won’t be too far off, as bots augment and replace traditional customer services. And that was in the very early days when virtual assistants and chatbots were starting to scope each other out as best buddies in the race for digital engagement.

More recent research shows it costs $8,800 to hire and train agents, while poor service results in around $62 billion in lost sales in the U.S. alone. But the AI need not replace the agents, with the main bot handling more tasks while giving agents useful information immediately to help handle the more complex queries, improve the quality and speed of their interactions.

And as bots as agents work together in most call centres or contact environments, both customers and workers will see the true value of their digital services. That’s vital as consumers and customers expect better customer service, merchant research highlighting that “The number-one business priority in 2019 is customer experience (CX), with 33% of eCommerce companies putting this first, behind even content marketing at 26%, and brand awareness and launching new products and services.”

If that doesn’t focus the mind of customer-focused businesses who are still on the fence about chatbots, then they will soon have a serious problem. And risk being outpaced by those with high-quality, easy to evolve cloud chatbot services that can make them more attractive to customers and able to grow business through positive brand awareness and word of mouth.

Chats Are Getting Broader and Deeper with Greater Understanding

The text or speech option broadens the power of chatbots and on-the-fly translation can give even a tiny business a global reach, all while bots remain easy to build and improve. Come 2020, the boxes are firmly ticked for chatbot efficiency, success and proven in the market.

Now, the debate is moving further with chatbots needing to offer broader conversations, focused on greyer areas of the chat. For example, how will your banking or financial services bot respond to, “is there a way to earn more long term with my $1,000 in my savings account?” The key triggers are in the question, the bot needs to be smart enough to pick up on the cues and offer a range of suitable options based on the amount and term.

As people have broader conversations with bots, they will expect to be able to pick up a conversation where they left off, or for the bot to remember the outcome of previous chats. That could be a conversation about a pizza order that started at home on Alexa but the kids’ changed their mind and now you’re on your Messenger app as you walk down to collect it. The more a bot knows about the customer and their interactions, the better the service will be.

Finally, bots through AI can deliver a better understanding of people’s needs, especially when it comes to medical, legal, banking and other sensitive areas. The rise of natural language processing (NLP) and Natural Language Understanding (NLU) help bots follow the pattern of a conversation and can pick up on stress points.

Let’s say someone needs an emergency overdraft and is using the virtual bot because they are too embarrassed to talk to the customer care helpline. Assuming the customer isn’t a serious credit risk, and the bot either through text analysis or voice stress senses some real need, could the bot override the usual “computer says no” expectation? Delivering a positive result, while highlighting the usual just this once caveats and repayment warnings, could build the customer relationship. Beyond that, can the bot check in to make sure the customer solved their problem, and remind them of repayments, building a degree of trust beyond the usual form letters or emails.