Dr. David Naylor is the founder and director of artificial intelligence (AI) company Humanotics.
Customer experience improvements are touted by the industry’s Great and Good as one of the most important areas where Artificial Intelligence will have an impact. Look at any Forrester Research paper or visit any solution vendor website to find the relevant whitepaper. You could easily think that if you haven’t got a chatbot programme up and running you are behind the curve. It’s important to remember that this market is still in its infancy and it’s wise not to get caught up in the frenzy – bleeding-edge is so named for a reason. While chatbots are often the place that my conversations start with clients and prospects, raising awareness of the potential to drive performance in other areas of the contact centre often creates as much excitement.
So what are the applications that I feel hold the most potential, we most frequently talk about and how do I advise clients to move forward?
Augmented Agent Assistance. This is a good place to start since it addresses the challenge that all customer service operations have – delivering effective support to agents during live interactions. We are talking about the tools given to an agent to support better interactions with customers such as suggested responses as part of Webchats. Also, they can simply provide the agent with a more intuitive way to access information in the knowledgebase. It’s worth noting that these suggestion tools can work on voice interactions too but the real-time processing power required frequently makes them cost-prohibitive. Making knowledge access easier for agents through a chatbot interface has several benefits. Organisations tend to see an uplift in the use of the knowledgebase which leads to more consistent customer experiences and faster resolution. The quality of the knowledge also improves as questions and answers become more refined, eventually allowing the information to be delivered direct to the customer without agent intervention.
My advice to businesses looking to get started in this area is to ensure you have someone responsible for the curation of knowledge and don’t assume your old and inaccurate content will miraculously improve when you push it through the AI knowledge builder tools. In many cases you may just need to start from scratch but it’s worth it.
Quality Assurance and Performance Development. This is an up and coming application area where AI can deliver the insights that transform the way you drive performance improvements. Speech analytics tools are now capable of delivering results that make automated call evaluation and scoring a practical reality. This instantly frees your team leaders up to focus more time on coaching. Going one step further, the potential now exists for machine learning tools to analyse the characteristics of your best performing agents, in sales, customer satisfaction or other areas. You can therefore map the individual capability or knowledge gaps across the operation based on hundreds or thousands of calls per agent, not just the two or three per month that were manually evaluated in the past. Connect this analysis with your Learning Management System and you have closed the loop – you know where your agents need training, you can see when that has been delivered and you can review the hard evidence of the impact it has had. Even without the LMS tools, building a true picture of performance and capability with speech and AI tools, unlocks some of the most powerful insight that customer operations managers need today.
The first step on this journey is to ensure your speech analytics tools do an effective job of monitoring and evaluating calls – you do have speech analytics, right!? This structured data can then be fed into the machine learning stage. You won’t find packaged solutions in this space as yet but all the components are available which is where our expertise is unparalleled.
Forecasting and planning. Predicting when staff will be sick? Yes, that’s one application but the whole process from long-term planning, through to forecasting and scheduling is an area that AI can support. For instance, marketing models are using AI to predict campaign performance so, if we know how many sales we are going to make on a particular campaign, why not predict how many service contacts we will receive relating to that campaign from customers chasing their order, for example? And when will those contacts arrive if the campaign lands on a Tuesday morning in February when the weather is unseasonably cold? Don’t forget, of course, the fact that you have 20 new agents who are less experienced and likely to have a lower first contact resolution rate. Now, maybe not all those variables will impact the campaign but at the moment it is hard to know exactly what does influence contact rates and when you have 20 campaigns running simultaneously, almost impossible to track. Most service operations would not be that complicated but most forecasting is still done based as much on the experience of the planning team (aka gut feel and spreadsheets).
Ask yourself, what would a 10% improvement in forecast accuracy be worth to your operation? That’s where I suggest clients start and it’s then clear what potential investment they can afford to make using machine learning tools to develop a more accurate approach. At this stage, Workforce Management tool providers have not addressed this, so be wary of their claims.
One final piece of advice that applies to all these areas is to start small, build something that works – a Minimum Viable Product – and then go for incremental gains. This is essential when tools and solutions are developing fast in a fluid market. Avoid locking yourself into one architecture or vendor too early or for too long as the landscape will look different in 12 months. But, do go for it – these AI applications offer the contact centre some great potential benefits to go beyond the chatbot.
Want to know more? Take Dr. Naylor’s masterclass on AI in the contact centre. Read our Ask The Expert on AI in grocery stores replenishment or our analysis on the future of AI. Follow these top 10 AI experts for regular updates.