Balancing AI and Live Agents
The modern contact center must strike a delicate balance with its resources. Economic pressures for costs of goods and services have not leveled off, and even though jobs reports in the US remain strong, concerns of inflation and recession are not abated. Supply chain disruptions resulting from COVID shutdowns still ripple through global industries, impacting product availability and delivery times. These factors are driving contact volumes up as fast as the wages and operating expenses needed for agents to handle those interactions. It’s more important than ever for the executives who oversee the customer journey to take advantage of new technologies to optimize their service platforms.
Despite the rush to automate, contact centers are equally pressed to maintain a focus on their agents. While automated workflows and generative AI have huge benefits and the potential to replace many of our agents, there are still cases where human talent is needed. MotionCX can help on both fronts.
Where to begin?
Perhaps the greatest challenge in implementing AI-based technologies is finding the starting point where there are so many great opportunities. The temptation to “boil the ocean” and go for it all at once must be avoided. Projects need to be planned in achievable steps. Running in too many different directions at once may result in a lack of resources, or projects that are not completed as priorities shift.
Another pitfall in project planning is beginning with something too advanced. Generative AI relies on accurate language models, and in many support scenarios there are industry- or company-specific terms and processes that must be accommodated. Machine Learning (ML) is also an iterative process, starting with basic steps and data sets that are built upon through time and repetition.
Crawl, walk, run.
The MotionCX team are experts in contact centers. Our experience with both the technologies and operating standards common to customer service across industries helps guide our clients to find the opportunities that will have the largest impacts in the shortest time.
Our measured “Crawl, Walk, Run!” approach starts with short-term achievable goals and the establishment of baseline KPIs.
In the Crawl stage, the applications are generally assistive to the internal staff. While this may have a customer-facing component, such as an Interactive Virtual Agent (IVA) verifying a customer’s identity before handing the interaction off to a live agent, the general scope is to collect information. The savings come from reduced time in a live agent interaction, and improved accuracy of information.
When we Walk, we are adding automation that leads to incident resolution. The automation may be internal, such as creating and assigning a series of follow-up tasks, including more timeline alerts and status notifications for both the users and the customer.
Running harnesses the potential of generative AI to guide conversations and capture the customer’s intent, and machine learning enables the system to understand the appropriate course of action to resolve incidents based on specific criteria. The IVA in Run mode is almost fully autonomous, and the live agents serve in specialized roles as either a QA analyst & supervisor over the AI processes, or in customer-facing capacity as a higher tier of support to provide white-glove services and to handle more delicate customer experiences that benefit from a personalized touch.