Artificial intelligence (AI) has unprecedented potential in revolutionising business practice and operations. The transformational power of intelligent, automated processes is rapidly ushering in a new age of productivity and profitability across numerous industries. This is particularly true in customer services.
IDC data shows that within Europe, 59% of this business function are planning to use AI in CX and 27% are evaluating to do the same. These stats are much higher than in any other business departments across the board.
This makes sense, as AI has the potential to solve the age old question of how to drive long-term, sustainable increases in efficiency, whilst maintaining and even elevating customer satisfaction, thereby stimulating greater engagement with customers and business growth.
It’s becoming clear that self-service is a critical differentiator in effective customer support. However, there is one challenge which remains; businesses need definitive steps on how to best implement AI technologies and what expectations they should have regarding timelines, output and return on investment. In addition, it is no secret that some businesses are dissatisfied with the results of having implemented AI within their customer service models, but this rarely has anything to do with the capabilities of the AI itself, more how it was deployed and how it is being used.
The AI implementation challenge
For business leaders looking to integrate AI into their customer service programmes, it might come as a surprise to learn that the most prominent reason behind disappointing results, is entirely down to a lack of preparation on their part. Often, those that claim that results don’t meet their expectations, have failed to take the time to research the challenges experienced by other businesses making similar investments.
It has been true for businesses across every vertical industry; new technology is investigated, purchased, installed, initiated and then the widely held expectation is that it will just work. However, this ignores the fundamental basis of AI technology, which is that the system learns how to manage situations based upon data and experience.
In reality, for an AI chatbot to be able to answer any given customer query satisfactorily, it requires a lot of background data. To provide context, this would (at the very least) include establishing several hundred different ways that a particular query might be asked by a customer, before even looking into providing the numerous response pathways available to the operator attempting to resolve the issue.
This is often cited as an obstacle for new implementers, who don’t have existing data at hand to pre-train the system. However, this isn’t the end by a long shot. Without the necessary training data, some organisations have elected to train chatbots on the job via machine learning. This is certainly the right approach to making AI work for a specific business need, however, it cannot be done in isolation and there must be a strong connection and feedback loop established with the wider support and service functions of the business.
This will act to improve outputs from the AI and ensure active learning to reduce reliance upon human operators. For those that don’t recognise the need to integrate all of their services, there’s only one conclusion; a basic chatbot function without self-improvement capabilities which cannot connect with human teams in the case of unresolved queries, which will undoubtedly lead to customer frustration and a negative brand experience.
What you should be doing
So what should businesses be doing to make sure their AI implementations are a success?
The first step requires establishing the core objectives; namely reducing pressure on the existing call centre and services operations and improving the customer experience through the modern technologies that consumers are coming to expect as a minimum offering.
Second, think big but start small. Expecting to replace customer advisors with virtual assistants in the short term is unrealistic. There are many competencies to conquer before that. However, it is likely that at some future point the number of advisors will reduce. Simple customer engagement will mainly be addressed through virtual assistants. While live assistance will be optimised through intelligent AI support and smart routing.
Thinking big and imagining how things could unfold over a period of time is certainly better than approaching anything as transformational as AI in piecemeal fashion. But having grasped the opportunity, it is then crucial to focus on your roadmap and the place from which you intend to start on that journey. For example, it would be advisable to start with online chat first, so that data can be collected from these interactions, before you move onto a machine learning chatbot implementation for that seamless transfer from self- to assisted-customer service.
Finally, set smart expectations, which grow over time. Those that pace their expectations for AI, to grow as the solution acquires more contextual data and improves its own output potential will realise its true power in driving efficiency and productivity. Learning technologies don’t work out of the box, and due to the privacy requirements around company data, the only way to train a machine is after installation inside the network. The most effective strategy for implementation involves preparation and analysis of the quick wins and big opportunities that chatbots can seize on in the early stages, before progressing to larger prospects at a more advanced stage.
Invest in the future
AI will completely revolutionise the world of business in the 21st century and the anticipation around AI enhanced applications, particularly customer services, is entirely justified. The reason why some are still trying to achieve optimal results rather than already reaping the rewards of intelligent automated services, is simply because they haven’t invested in understanding the technology and it’s needs before integrating into the workflow.
With rapid advancements in technology and the increasing automation of simpler roles in wider enterprise, it’s inevitable that basic customer services will become largely managed by AI platforms in the near future. However, to achieve the promised benefits of AI, businesses have to set smart goals in the short-, mid- and long-term. They must also invest time and resources into developing a solution that is tailored to both their business and their customers, with constant data input from across the entire customer service function, to ensure growth and relevance as customer demands evolve.
By capturing quick wins and building the complexity of AI chatbots on the job, it’s possible to achieve a new level of efficiency and customer satisfaction, whilst creating a significant competitive advantage that will enable greater success for businesses navigating a crowded and increasingly digital marketplace.
