Search

The 2nd Wave of AI in Customer Service: Collaborative Intelligence

For well over a century the question of man vs machine was all the rage. Competitions ensued as people were drawn to the speed, output, and intellect of machines yet terrified by the specter of being replaced by one of them. It came to a pivotal moment in 1996 when Gary Kasparov took on IBM Deep Blue in chess and won. Man prevailed…at least for a short while.


The very next year Kasparov lost to Deep Blue in a rematch and he subsequently lost or tied with other computers thereafter. But instead of focusing his efforts in trying to defeat the machine he transmuted man vs. machine to what can humans and computers achieve together. Thus he started ‘advanced chess’ which brought together humans and computers on the same team to raise the level of play.


Humans + machine = better together

Fast forward to 2005 and the PAL/CSS tournament where the unthinkable happened. Two average chess players using AI beat a team of three chess grandmasters using way more powerful AI. How? They were experts at collaborating with computers. They knew when to rely on their own thinking and intuition and when to rely on the machine’s advice. This showed that combining human and machine intelligence can yield significantly better outcomes than either can achieve alone. This is called collaborative intelligence and Jack Johnson's "Better Together" would be the theme song.


From navigating traffic dynamically through changing environments with Waze, to using suggested comments to Linkedin posts and messages, to cruising with autopilot on Teslas, collaborative intelligence is increasingly becoming mainstream in our every day lives. In the enterprise, however, collaborative intelligence is just taking off – and customer service is the low hanging fruit.


Customer service has reached its breaking point

According to results from Forrester’s annual CX index (that surveys hundreds of thousands of consumers) customer service is the top driver for customer experience in most industries. For many companies when chatbots, IVR and other self-service touchpoints fail, customers will seek the help of a live agent. And as the last resort for customers the agent is the determining factor for those “make it or break it” moments that will decide if a brand shines or fails. Therefore focusing on improving the agent experience is a no brainer.

But the job of the agent is harder than ever before:

  • Call times are longer – Average handle time (AHT), before the pandemic, was up 12% as calls are more complex and take longer to resolve according to Concentrix. This number has increased based on recent conversations with customer service executives further exacerbated by secular forces including supply chain shortages, flight cancellations, inflation, labor shortages, etc.

  • The Great Resignation has come to contact centers - Attrition rates are at an all time high at 58%. Some companies report over100% in attrition rates. Pandemic/post pandemic rage is still very much present destroying the agent experience. Contact center executives have told me that some of their agents have expressed they would rather work in fast food, where pay is slightly less but they don’t get yelled at by frustrated customers.

  • More agent knowledge is needed – Today agents have to be more knowledgeable than Google as a majority of customers turn to google searches at the onset of their journey to resolve their issue. This means they will be better informed (or even mislead) from ingesting content (discussion board, FAQ, etc) before engaging with agents. The agents need to be prepared and deliver value in these instances with insights and knowledge.

The stars are aligning for collaborative intelligence

There is some good news though. The market and technology are ready. According to McKinsey’s 2022 State of Customer Care Survey the top three priorities (for companies) over the next 12 to 24 months will be retaining and developing the best people, driving a simplified customer experience while reducing call volumes and costs, and building their digital care and advanced analytics ecosystems.

Many of the technologies (speech recognition, transcription, natural language) that support collaborative intelligence for agents has improved dramatically with the use of GPUs, refined algorithms, and deep neural nets. .Overall collaborative intelligence is poised to make the leap from the sci-fi world of Minority Report and Ironman to an actual applied technology in the contact center.

Here are five ways collaborative intelligence will help contact centers today

The combining of humans and AI in contact centers goes beyond cool to actual business impact as collaborative intelligence can help:

  1. Improving first contact resolution (FCR) and decreasing AHT. Imagine the start of a conversation between a customer and an agent. As the customer describes her issue (over the phone or via chat) the agent’s screen dynamically populates (on its own) the specific information and right workflow to resolve the customer’s issue. The agent did not have to press any button on the keyboard or mouse. Then as the conversation continues the tool listens in and provides real-time dialogue suggestions, behavioral cues, and knowledge surfacing (articles, forms, videos, etc) to the agent. Streamlining information lookup and mashing that up with the right workflow for resolution dramatically improves FCR while decreasing AHT. In fact, a Minerva CQ client decreased AHT by almost 50% and increased FCR by more than 10%.

  2. Reducing agent stress and workload to improve agent experience (AX). Collaborative intelligence can bring down agent stress levels significantly as agents react to changing customer sentiment in the call simultaneous to overcoming knowledge gaps. This is where collaborative intelligence shines. Contextual knowledge surfacing decreases the training time or eliminates it entirely for the agent in some cases. For example, agents will always be up to date on the latest, device support updates or mobile app changes without having to go through training, or constantly searching through knowledge bases or paper documentation.

  3. Accelerating agent onboarding time. The typical agent onboarding time is 4-12 weeks. This obviously becomes more complex with WFH agent operations. The industry has seen 25%-33% reductions in time-to-efficiency for new agents through the use of collaborative intelligence without any disruption to the call flow. Most of this is driven by the prescriptive advice and easier navigation that is provided by collaborative intelligence. A Minerva CQ client forecasts 75% reduction in agent onboarding time with collaborative intelligence.

  4. Identifying struggling agents and coaching them to success faster. Each conversation has multiple data points that collaborative intelligence draws from. This level of granularity provides more insights for team leads and supervisors (or even automated QA tools) to identify and coach the lowest quartile (those that may not be a company fit in the early days) to be more effective in their roles faster. This also helps mitigate the cost impacts of high attrition and long, expensive hiring cycles.

  5. Extracting business intelligence and VOC. Collaborative intelligence provides for unique, in-the-moment insights about all call and chat types. This will include an understanding of optimal call (and chat) path baselines and agent adherence, real-time FCR, NPS, and CSAT in aggregate as well as for individual calls and chats, as well as data-driven insights into how call (and chat) paths and business processes might need to change, based on prioritization of KPIs and certain commercial levers. This provides a beacon for managers to stay ahead of events and changes to help inform agents and prepare them for upcoming rough patches.

The death of the contact center is not happening anytime soon much to the chagrin of those who prognosticated its demise. But what we are seeing is the intertwined evolution of people and AI in customer service as customer needs become more complex and the agent job becomes harder and more specialized. We are just at the beginning of this developing area of collaborative intelligence in the enterprise but over the next decade this will be where many of the “wow’ moments will come from as AI is finally being used to enrich the customer-agent engagement as opposed to preventing it through deflection, self-service, and automation.