Building a bot that people actually like
It’s happened to most online shoppers – you’re browsing for an item, and suddenly, a chat window pops up with an overly enthusiastic yet obviously automated bot asking if you need help. You roll your eyes, annoyed at the interruption, and quickly click to close the window. It’s a frustrating experience that highlights the challenges many companies face in implementing virtual assistants and chatbots that customers actually want to engage with.
Recent surveys reveal consumer resistance to the bots invading customer service channels. Despite heavy investment in the technology, satisfaction rates remain low as users find typical chatbots lack empathy and conversational ability. A survey undertaken by market analysts Gartner midway through last year showed just 8 per cent of customers used a chatbot in their most recent customer service experience, with only a quarter of them saying they would use that chatbot again.
This highlights the need for organisations to implement bots thoughtfully, with the customer experience at the forefront. The goal should be a seamless human-bot collaboration that enhances, rather than detracts from, customer interactions.
Understanding user expectations
A key step is understanding user expectations and reluctance. Interacting with a robot versus a human agent provokes different reactions.
TSA Group IT Specialist Software Engineer Tim Reynolds says bots that lack human traits like warmth, personality and humour tend to underperform. Experimenting with giving bots character and personal flair helps humanise them. Mr Reynolds, however, says it’s equally important to set accurate expectations by making it clear users are interacting with an AI system, not a real person.
“It becomes a bit of a balancing act and one of the bigger challenges is determining how long to let the chatbot prompt the customer before escalating to a human agent,” he says.
“That’s probably where a lot of reluctance and negative sentiment comes into play, with people who have been stuck in interactive voice response queues when they just want to talk to someone. But there are all sorts of smarts that you can put in place to try and get information upfront or identify if it is a sensitive case, and in that case, the customer can skip the bot and go straight through to talking to a human agent.”
Bots pose risks too
While chatbots offer many advantages, implementing them also comes with significant risks that companies must mitigate. Mr Reynolds says a major concern is the potential for bots to provide customers with incorrect or misleading information. He says without human oversight and intervention, bots may make errors that negatively impact the user experience.
“Giving incorrect information can also negatively impact the company’s brand or reputation,” Mr Reynolds says.
“And if it’s misconfigured and customers are spending too long with the automated bot before being passed on to a human, then a higher percentage of calls will land with human agents after the customer’s sentiment has already dropped. In that case, agents then have a worse experience than if the customer had been transferred straight through or had a shorter time with the bot before being escalated.”
Mr Reynolds says companies should put strict governance and quality control processes in place to ensure bots are regularly tested, monitored and updated with accurate information. He says rigorous quality assurance protocols help minimise the chances of bots propagating bad data that confuses or frustrates users.
“Any customer-facing AI implementations need to have guardrails on them and need to be heavily tested to ensure that the responses that they are giving are relevant and suitable, and not wildly inaccurate,” Mr Reynold says.
“And there will always need to be someone monitoring what the outputs of bots are and ensuring that it is actually correct. Bots won’t necessarily replace a role, but people’s roles will definitely modify and change over time into more of a maintenance and review sort of role.”
“Another important consideration is around business processes and being able to document those processes. It’s far easier if you already have chat or voice services running to be able to analyse and understand what customers are calling for. What may work for one line of business may not necessarily work for another line of business, and they really indeed to be tailored to the specific lines of business.”
Data security vulnerabilities
Along with quality control, Mr Reynolds says security is another critical consideration when deploying chatbots. Bots that interact directly with customers must adhere to strict cybersecurity protocols. There is always the risk of data breaches if proper protections are not in place.
Chatbots with access to sensitive customer information could expose organisations to compliance violations and legal repercussions if hacked. Companies need to perform thorough risk assessments, enforce access controls, encrypt data end-to-end and establish contingency plans for potential bot security incidents or outages.
With deliberate strategies to provide reliable information and safeguard user data, companies can contain the unique risks posed by chatbots. This careful approach helps build trust and ensures customers feel comfortable engaging with AI.
Determining if a chatbot is right for your business
While chatbots offer immense potential, they aren’t necessarily the right solution for every business.
TSA Group IT Specialist Software Engineer Christian Sesta says companies should carefully evaluate whether their customer engagement needs are best met through human agents or automation. Bots excel at handling high volumes of simple, repetitive queries but may falter when it comes to building relationships or providing personalised service.
“It really comes down to the business’ need,” Mr Sesta says.
“Not all businesses need chatbots because they may not be able to provide all the necessary information.”
Mr Sesta says companies should consider if a bot aligns with their brand identity and customer service philosophy.
“Implementing bots for the sake of trendiness without a clear strategic purpose will likely deliver disappointing results,” he says.
“With careful analysis, organisations can determine if integrating a thoughtful, well-designed chatbot makes sense for optimising their customer experience.”
TSA are Australia’s market leading specialists in CX Consultancy and Contact Centre Services. We are passionate about revolutionising the way brands connect with Australians. How? By combining our local expertise with the most sophisticated customer experience technology on earth, and delivering with an expert team of customer service consultants who know exactly how to help brands care for their customers.