What are the key challenges currently facing the UK IT and telecoms Channel in adopting AI in contact centres?
The main AI adoption barriers in contact centres are twofold.
Historically, the top challenge has been customer sentiment towards the experiences smart technologies provide. Customers haven’t responded well to chatbots that are only able to give formulaic answers to their queries, resulting in low confidence in using AI for customer-facing tasks over the last five to 10 years. This hesitance was especially high following the pandemic push towards digital transformation, with many businesses taking a dim view of AI after hurriedly implementing automated service solutions and finding they couldn’t match the depth of human support.
However, In the past 12 months, we’ve seen a huge revolution in AI sophistication. Generative models can now deliver incredibly positive experiences when used correctly. With our AI-enabled support line, for instance, there was a learning curve for customers to realise that they can ask questions naturally and converse with AI tech, instead of using keywords as they would with interactive voice response (IVR) systems. So, the experience quality is now much better and more nuanced, but we need to work on helping customers understand that.
Trust is the second barrier. As shown by recent concern around AI analysis of customer calls without pre-approval, there is a need for stringent compliance and adherence to responsible usage policies, such as gaining explicit customer consent. At the same time, there is also a high burden of trust for companies, particularly small-to-medium-sized businesses.
On the service side, many are worried that introducing AI at critical points of contact could have significant ramifications for customer satisfaction, loyalty, and long-term organisational success if interactions go wrong. Meanwhile, developing tools is currently driving high costs for vendors that many are swallowing to make products available at a consumable price.
For both of these challenges, time will prove AI’s value and eventually bring down the cost of tech.
In which areas is the technology currently making a big difference in contact centres?
We’ve had access to what has been defined as AI for decades. But the generative shift has brought a step change in what it can do at a practical level for every function, including customer service. What the latest breed of machine learning tools are especially good at is analysing large volumes of information, identifying patterns, and producing novel insights and recommendations. All of which makes it the perfect fit for guiding agents to take the next best action for each customer.
At the same time, smart and well-considered automation is bolstering the efficiency of routine service tasks and easing pressure on often overstretched agents. When deployed as the first line of support for more straightforward issues — such as troubleshooting digital devices or updating account details — AI assistants help to significantly curb delays on the customer side (no more waiting endlessly in call queues) and liberate time for agents that can be reallocated to tackling more complex challenges.
Additionally, implementing more pockets of self-service support is freeing companies from ongoing concerns over interaction costs. Every conversation costs money and, for many SMBs, this means a restrictive focus on trying to keep interaction numbers down.
With virtual agents taking some of the load, however, firms have the scope to lift these limitations on service availability, allowing customers to reach out as often as they need and consistently receive a high level of AI-fuelled personalisation. In fact, future service performance may be judged against how frequently customers contact companies via multiple channels as a measure of their connection with the company.
How can AI improve customer experience and build trust?
Previously, most contact centres would only have the bandwidth to analyse a few customer interactions and hope that they could provide an indication of overall service quality. Thanks to developments in machine learning, they can run in-depth evaluations at scale, gaining a transparent picture of every conversation, as well as recommendations on the opportunities available to improve experiences. In short, they can assist agents by equipping them with the data needed to make better decisions and earn customer trust through elevated service quality.
Of course, we have to recognise the persistent challenge with AI and trust. AI itself doesn’t necessarily garner trust. Instead, we need to build trust in AI by demonstrating it can be leveraged reliably and effectively. As mentioned, one element of that is establishing rigorous compliance procedures for human users — but there is also a requirement for self-governance. To trust that AI won’t risk damage to customer relations, companies must apply the same level of scrutiny to AI assistants as they do to agents: including instructing smart tools to monitor for and flag any potential errors in their own activity.
How can it provide service agents with greater insight into customer behaviour and sentiment?
Beyond enabling agents to track, understand, and fine-tune their performance, analysing the sentiment of customer interactions across organisations offers a richer basis for assessing customer trends — covering highlights and pain points — and re-shaping the company’s entire approach to drive stronger experiences and, in turn, greater trust.
For example, part of that might be spotting and phasing out ways of handling issues that consistently result in negative emotions and low satisfaction. Conversely, it may involve finding and amplifying communication styles that leave customers feeling positive about the agent and company.
As with any form of analysis, evaluation of customer sentiment comes into its own once firms have sufficient long-term data to map customer interactions and develop a complete understanding of their habits and needs. This information is the ideal basis for building unique profiles that define exactly how customers want to be served and ensure every experience fits their specific requirements and communication preferences. Moreover, running continual analysis will also generate further insights into personalisation hits and misses that can be tapped to optimise interactions for even better results over time.
How can it improve decision-making and make agents more engaged, empowered, efficient and productive?
This brings us back to the value of AI in steering agents to the next best action for their customers. By nature, service jobs involve managing tough problems and dissatisfied customers who want their issues resolved quickly. Having access to AI tech is vital for relieving much of this burden, allowing them to achieve their ultimate end goal: customer satisfaction. This ultimately fosters an environment where teams feel more comfortable, engaged, and empowered to do their job well.
In terms of productivity, process automation also enables agents to maximise their expertise. Rather than manually wading through labour-intensive admin, service specialists can focus their unique knowledge on solving higher volumes of more complex problems – simultaneously clearing bottlenecks and increasing job satisfaction.
How will the use of AI in contact centres evolve over the next 12 months? What new applications are we likely to see the technology being deployed for?
The greatest evolution will be mass integration of AI into contact centre teams. We are quickly approaching the point where smart assistants functioning alongside human agents and driving a sizeable share of operations will become part of the general workforce — essentially acting as employees.
This change will raise difficult questions about how companies can ensure tech usage doesn’t tip the service pendulum too far towards mechanisation. For many companies, however, AI will play a crucial role in helping to address long-running hiring issues by filling staffing gaps, such as up-scaling team capacity during the festive season when workloads tend to spike.