Lately, a few organizations in South-East Asia have started using chatbots to respond to customer service enquiries. Marketed as their own AI (Artificial Intelligence) entities, these chatbots are currently being used by these organizations to answer Tier 1 (or Level 1) customer service issues, and handle the mundane tasks of data gathering and insights building on the issues at hand.
More and more organizations seem to be getting on the bandwagon – according to a chatbots developer (Chatbots Retail, eCommerce, Banking & Healthcare 2017-2022, Juniper Research, July 2017), these chatbots are expected to cut business costs by US$8 billion by 2022.
And here’s why – one of the main draws of chatbots for organizations has been the potential of leaving mundane tasks to machines – this in turn:
- Reduces manpower costs;
- Frees up talent resources to focus on higher value work;
- Amps up productivity, and
- Can improve workflow processes.
There’s also the service aspect of things. Chatbots don’t need to eat, sleep or take breaks, but customer service teams do. Letting a chatbot take the lead on round-the-clock enquiries allows a business to maintain a 24x7x365 presence at a fraction of the cost.
Is this really the case though?
Most chatbots adopt an opt-in approach; they only respond to user prompts. Chatbot conversations are initiated by people, and people only do so when they are ready to be engaged. While the younger generation may be more than comfortable to engage with a machine, many older people (the ones with the money) have their doubts about responding to a computer program that, in their view, is hackable.
With concerns of data theft, and the ability of sophisticated AI-based chatbots to make intelligent recommendations by remembering user inputs and cross-referencing this to existing data on a customer, those of the older generation may not be too keen to engage with a chatbot even for their Tier 1 customer service experience.
Add to this headache is the issue that so far, success rates have been mixed. Where users have been willing to try to interact with chatbots, the observation has been that they have a lower tolerance for mistakes by machine vs man. When a chatbot cannot understand what a user is saying or forgets what they said a few speech bubbles ago, frustration is inevitable.
So even if it does save a lot of time and resources, nothing beats investing in real, human customer service support people. Many organizations in South-east Asia still have not built sufficient strength in their brand experience for them to be relying on chatbots, and this is an important first step.
It will take a few more years for organizations in the region to understand the interactions their brands have with their audience before investing into chatbots reliably.
This is because the real problem isn’t technical in nature – it’s conversational. It’s about understanding the interactions the audience already has with their brand and finding a sweet spot where the chatbot interface can improve, not interrupt, that process.
