Firms in all industries see artificial intelligence (AI) and automation as a means to improve operational quality and customer experience (CX), reduce costs and increase margins. Customer service is a good place to employ such technology as it’s both a major expenditure and a driver of customer experience. AI and automation can help at every step of the customer service journey.
For instance, conversational AI or chatbots are appealing to organisations looking to offer customer service outside of business hours without requiring extra staff. By removing repetitive, low-value tasks from customer service agents, automation lets them focus on where their human touch adds the most value.
This is how hotel chain Hyatt uses a virtual assistant for parts of its reservations journey, taking away the “mindless” draining tasks – such as authenticating customers or gathering their travel dates and destination – before transferring the call with all the relevant context to an agent who can focus on the emotional component of the selling. Finally, the ability of the agent to connect to the customer is a fundamental driver of call quality.
Risk of having a narrow view of customer journey
As artificial intelligence and automation technology matures and becomes more affordable, it’s tempting to rush to deploy it, but this can result in common pitfalls that prevent a business from reaching the impact they expected – or worse, that hurt the customer experience.
This can happen if there is a lack of insight into the end-to-end customer service journey. A customer service journey involves multiple digital and non-digital channels, as customers typically try to self-serve before reaching out to a customer service representative. CX professionals who target specific moments or touchpoints in the journey without a clear and complete picture fight a losing battle.
“Suppliers of automation and AI services say their main challenge with customers is managing expectations about what technology can and cannot do”
If call centres are overwhelmed because customers can’t find the information they’re looking for on the company website, the first action should be to fix the website rather than deploying a chatbot to answer those requests.
Misjudging technology capabilities is another problem organisations can encounter. Suppliers that operate in the automation and AI market agree that their main challenge with customers is managing expectations about what technology can and cannot do. The best return on investment comes from high-volume, simple use cases that can be answered without the need to hand over the interaction to a human agent.
But when companies get it wrong, they create frustrating experiences for customers. Conversational AI or chatbots that lock customers into dialogues, redirecting them from one unhelpful tool to another, are a common illustration of such misjudgment.
It is also worth noting that these technologies should not be considered as replacements for humans. The Covid-19 pandemic accelerated the rate at which machines took on human jobs. Workforce reduction is still considered a potential benefit of AI, but while AI is transforming customer service, it won’t replace human agents. Human representatives are necessary for highly emotional or complex cases where customers seek human interaction.
Blending human and AI customer service
To reap the benefits of AI and automation, you must identify their right place and role in your customer service journey. Forrester breaks this down into six steps.
The starting point is to map out the service blueprint of your customer service.
Opportunities for AI and automation often reside in the backstage of the experience. Start mapping the visible part of the customer service journey, including before and after interacting with an agent. Then add the invisible layers of the experience – the technology and processes that enable or hinder steps of the journey. Once you have a complete service blueprint, highlight pain points for all actors that are part of it.
Forrester’s second tip is to apply the five whys technique, which iteratively drills down into a problem to identify the root cause. Use the five whys technique to perform a deep root-cause analysis of your pain points and assess if you really need artificial intelligence or automation.
“Prioritise opportunities that benefit customers and employees – for each opportunity identified, evaluate who will benefit and then prioritise those that benefit both employees and customers”
A UK police force, for instance, used service design to discover that a large number of 999 calls were mere requests for information and were hindering call handlers from helping citizens truly in need. Additional research showed that fragmented processes and workarounds for existing blockers were reducing efficiency and visibility of the process for citizens. The solutions focused on better access to information rather than automation.
Forrester’s third tip is to prioritise opportunities that benefit both customers and employees. For each opportunity identified, evaluate who will benefit and then prioritise those that benefit both employees and customers.
One company that tried this approach is BT, which used AI to improve customer service by focusing its field engineers on the right job at the right time. Instead of using the workforce management system to assign jobs only to local engineers, the company used fuzzy logic to enable field engineers to cross regional borders, resulting in better service, increased productivity, reduced travel costs and improvements in employee well-being.
The fourth recommendation is to identify the right technology. The terms cover such a broad range of capabilities that the help of a subject matter expert might be needed to define what’s right for a particular company, depending on its existing systems and use cases. However, there’s one technology that’s fundamental for the contact centre – speech recognition.
Dutch telco KPN, for instance, used speech recognition to reduce its average hold time by 30 seconds per call and increased its net promoter score by 17 points. The customer begins by stating in their own words why they are calling. AI authenticates the caller, recognises their intent, and automatically answers or routes them to the right agent. When it does so, it pulls out the customer’s details and call history and transcribes their own words so the agent immediately has the right context.
Next is to make emotional connections beyond empathy part of the requirements of the AI and automation project. Subtle changes in words can improve customer experience by creating a positive emotional connection with customers. Content strategists know that what you say and how you say it can be a powerful differentiator. AI can generate emotionally engaging content, but you need to ensure its emotional capabilities.
For example, the team building Capital One’s AI assistant, Eno, focused on developing its emotional intelligence. As a result, the bank saw customers sending Eno gratitude messages after using it as if they had experienced a positive interaction with a real person.
Forrester’s final recommendation is to align success metrics for the end-to-end journey. Finding the right metrics to measure the return on investment of automation and AI requires a careful assessment. Avoid isolated metrics. Rather, define a set of journey-level quality metrics to measure impact correctly.
One story of woe is that of a telecoms company which had focused on reducing call times in its service centre when onboarding new fibre network customers. But when the firm partnered with McKinsey & Company to analyse journey data, it found that reducing call times caused more follow-up technician visits. This cost the firm between 10 and 20 times what it had saved by shortening call times.
This article is based on an excerpt of the Forrester report, “How AI and automation drive better customer service experiences”. Karine Cardona-Smits is a senior analyst at Forrester. Ian Jacobs is a principal analyst at Forrester.