One often-cited approach to improving the performance of contact centers and customer service agents is skills-based routing. This involves tagging data about the skills of individual agents – for example, languages spoken, training courses passed or the ability to handle well a particular type of call – and using a call-routing system to deliver calls to an extension where an agent with the requisite skills has signed in and is available. Identifying the required skills typically is done by an interactive voice response (IVR) system or perhaps through the number dialed by the caller; in the latter case, a high-value customer might call a special number and identify the issue by selecting among options in an IVR system. Either way, matching customers and their requirements with agents skilled in dealing with them is thought to increase the chance that the customer’s issue will be resolved efficiently in the first attempt.
IBM has been working with one of its customers to make skills-based routing more sophisticated to increase the chance of making the best match. The result is the IBM Real-Time Analytics Matching Platform (RAMP). The logic behind the product is straightforward. Companies have lots of data about their customers in various systems: billing, CRM, ERP, customer data warehouses and others. RAMP has extraction tools that can take this data and build an analytics-based model of every customer. The same is true of data about agents, although it is held in different kinds of systems such as workforce management, quality monitoring or human resources and is part of contributing to what I call Agent Performance Management for which I have done extensive research including a recent benchmark. Using this data RAMP can build an analytics-based model of the agents’ skills and past performance. The system also can collect real-time operational data such as which agents are signed in, queue lengths, average call-handling time, targets for service level agreements (SLAs) and any updates specific to that day. Then RAMP combines the operational data with the customer and agent models to create a best match, also called an affinity score, between the caller and all signed-in agents. This happens in real time so the call-routing software can send the call to the agent most likely to deliver the best outcome. Should the most qualified agent be on another call, the system is smart enough to calculate whether to wait for that agent to come free or to pass the call to an available agent with a lower affinity score who might still deliver the desired outcome. The models are self-learning, so over time the matches continue to improve.
According to IBM, the results achieved by its initial customer are impressive, showing significantly increased customer retention rates. What is more those customers have remained customers for longer and over time have bought more from the company. For new user companies, getting these results will require investment. They will have to work with IBM to build the analytics models, develop the extractors to populate the models, build the affinity models and populate the real-time operational models before having these innovative routing capabilities in place.
My benchmark research into the use of technology in contact centers showed only a minority of companies having deployed advanced call-routing such as skills-based routing. However, companies that have deployed smart routing have seen customer satisfaction rates improve and realized business benefits by matching callers with the most qualified agent, even to the extent of routing callers to the agent they spoke to previously about an issue. Having an agent familiar with the customer and the type of call can reduce the lengths of calls, raise first-call-resolution rates and increase up-sales because the agent has a track record of completing more sales. I believe more companies could benefit from these capabilities and recommend that interested companies evaluate IBM’s highly innovative product for this purpose.
Let me know your thoughts or come and collaborate with me on Facebook, LinkedIn and Twitter.
Regards,
Richard Snow – VP & Global Research Director