Skip to content →

The path to AI-enabling your contact centre

The overall potential of AI within customer service has now been a hot topic for several years – but where does it deliver real benefits and what is the state of AI deployment? To find out, the latest ContactBabel Inner Circle Guide to AI, Chatbots and Machine Learning report, sponsored by Enghouse Interactive, surveyed both contact centre leaders and a cross-section of consumers.

The key finding is that AI is spreading, with a growing number of use cases identified by customer service teams. Just under half (48%) of contact centres say AI will be relevant to them with general agreement that it will affect contact centres of all sizes. While currently contact centres within the utilities, insurance and outsourcing sectors report the greatest use of AI – 35% of organisations say they are intending to implement it in some form within the next 12 months.

However, as with any new technology project, AI deployments need to be managed carefully to ensure long term success. This is particularly important given sensitivities from customers and concerns that it will reduce the need for agents, and thus lead to redundancies.

AI implementations therefore need to be clearly focused on areas where they can add most value to the customer, including providing fast access to routine information and freeing up agents to deal with more complex queries that require human skills such as empathy. Organisations therefore need to focus on the right use cases for their customers and then put in place a clear plan to deploy, monitor and improve AI implementations.

Exploring the potential use cases

The report identifies four broad areas in which AI is currently being used to drive improvements in customer contact:

  • Self-service: For example, using AI to understand what customers are aiming to do online and suggesting information or next actions, as well as providing 24/7 self-service and information, such as through chatbots.
  • Assisted service: Using real-time analytics on calls to improve the quality of interactions, agent behaviours and consequently customer outcomes at scale.
  • Interaction analytics: Using real-time data to predict customer actions and requirements, thus improving cross-selling/upselling opportunities as well as monitoring, quantifying and improving agent performance through post-call analytics.
  • Robotic Process Automation: Removing manual steps from processes, reducing administration and thus increasing productivity – for example by automating post-call wrap-up activities.

4 steps to follow for a successful AI deployment

Implementing AI can be a major project and the ContactBabel report outlines some key steps to follow to ensure it delivers results.

1. Identify the right use cases/metrics and set expectations

Start by looking at pain points in the customer journey that need to be improved. For example, do you want to avoid customers having to call/speak to an agent, or to automate security aspects of a call, so that the agent can focus on service from the start? Then consider if there is sufficient volume of data in order to train an AI system effectively, as well as looking for areas that are measurable so you can track the ROI of your AI project. In any new technology implementation there is a risk of failure so manage expectations carefully.

2. Underpin AI with knowledge and data

Having good quality knowledge and data is vital to success; with the lack of information being one of the biggest reasons AI fails. So, invest in collecting information into a single knowledge base that can be used by both AI and human agents. Keep tracking and updating this knowledge base to ensure it remains current, identifying and cleaning data so that it can be used to train AI effectively.

3. Be open with customers (and employees) and ensure you have a clear escalation path to humans

88% of customers believe AI should not be hidden from them, so be transparent with customers when it is being used. The same goes for employees who must be told why AI is coming in and how it will affect their role.

AI doesn’t always successfully resolve more complex queries, so make sure there’s a clear path to escalate to a human agent. 80% of contact centre leaders said that the lack of a smooth handoff led to abandoned self-service sessions for example.

4. Continually monitor performance and use analytics to identify new opportunities for AI

As with any new technology implementation you need to set clear metrics for success and continually monitor performance. What results are you actually achieving, and how can they be improved? Make sure these metrics link to your KPIs so that they are easily understandable at every level of the organisation. Once you have some AI success, look at how you can spread the benefits. For example, use analytics to understand where AI can automate new tasks, bringing greater value.

As the report outlines most contact centres are at the beginning of their AI journey, but companies increasingly see the potential benefits it can deliver. However, to deliver this success it is essential to recognise that AI’s main emphasis should be on improving the customer experience first and reducing costs second as part of a customer-centric strategy. In conclusion, ContactBabel therefore sets out a key rule to follow: “The use of AI should be focused on use cases where the AI does a better job than a human, whether that’s being quicker, more accurate, available 24/7 or able to see patterns in data that no person could see.”

To download the full report, click here.

Published in AI IVR/Self-Service Real-time speech analytics