Chatbots promise a lot to companies looking to minimise customer service costs and improve customer outcomes by getting to the right answer quicker. But the early implementations have revealed shortcomings that can drive customers crazy.
Already companies are employing AI-powered chatbots to respond to queries, help users find products, and troubleshoot product issues. Dominos, H&M, and RBS all use chatbots across very different industries for quite different use cases.
The success or otherwise of any chatbot deployment depends on a few critical factors. Invariably the best chatbot executions deliver these 4 things to customers:
1: Communicate in the channels your customers want
Already many companies have pigeonholed chatbots to online or in-app uses. Big mistake. Both run into limitations with usability. Apps might not get installed, get turned off or get uninstalled. And web-based chatbots invariably require the customer to be at their desktop.
AI-powered bots can be used to power other channels including SMS chatbots or voice chatbots. The best thing is, the AI learning gained through one channel can be applied directly to another as all are powered by the same database.
The key is to think through the use case and the customer journey, then decide what channel is best.
2: Allow customers to speak to a person if required
The day will come when chatbots get it 100% right. I think. Maybe we’ll never get there. Until then, there will always be a need for customers to get in touch with a ‘real person’.
To facilitate this, you need to have the functionality to allow customers to speak with a customer service representative if they aren’t getting a resolution.
In SMS or webchat, the way to do this is to cater for typed requests to ‘speak to someone’ by automatically transferring to a live customer care agent. Obviously, this requires resourcing on the other end.
On a voice call, this can be done two ways. Firstly, by transferring upon getting a direct request to speak to a person. Or secondly, by sensing increased stress or frustration in the customers’ intonation and pre-emptying a request to speak to someone with a call transfer.
3: Some degree of emotional intelligence
Chatbots ultimately need humanizing qualities to improve the ways in which they can help us. As we know, humans use words to get their point across, but rely also on tone, voice inflection and facial expressions too.
While technology will keep evolving to cater for these behaviours, a strong start has already been made. As we’ve mentioned voice chatbots already have the ability to sense changes in intonation.
Preliminary work on incorporating motion-sensing cameras in machine learning is already underway, too. It’s all about sensing frustration, anger, happiness, relief and combining these inputs with the customer’s words to better understand them and provide a timely resolution.
4: Save the customer time
This final point is super simple, but absolutely critical. Chatbots need to be better at answering questions and providing solutions than their alternative. Say, FAQ sections, people-powered chat, a customer support line or crowdsupport.
This could be down to the accuracy and refinement of the AI and database behind the chatbot. Tell-tale signs of shortcomings here are customers needing to repeat enquiries, or getting stuck in loops. From a customer’s perspective, both are entirely unacceptable.
But it may also be a product of the communication channel. Amongst the top complaints of chatbots is the fact that customers need to converse by typing – which means more time to complete the task. Having a chatbot that supports voice addresses this.
Are you serious about helping customers?
Too often the execution of a chatbot deployment is driven by the desire to cut costs. After all, it’s about having AI-powered auto-responses answering the queries that people used to.
However, if your company is serious about maintaining or improving customer service you’ll need your chatbots to give customers these four things.
For a free demo of Upwire’s drag and drop voice and SMS chatbot capability Contact Us.