Automation Vs. Expertise
Imagine for a moment that you have an unlimited budget for technology. And as long as we are dreaming, let’s also pretend that any system you put in place will get installed without a hitch and will function as advertised. Strategy comes pretty easy in a world like this— most of us would simply get the best of everything out there. If something is not a perfect fit for us just yet, we could always just stick it on a shelf and bring it back out when the time is right. OK, now let’s snap out of the unicorns and wizards fantasy land. In the real world, we not only have budgetary constraints, but we have to concern ourselves with limitations on technical resources, integration with existing systems, network capacity and loads of other considerations that make technology such a challenge. When obstacles and issues abound, strategy matters. But in the process of determining “which one,” how many of us ask the equally important question “how much” ?
That question may at first glance seem unusual. Most of us spend so much time trying to get the technology we need, we never consider that we may have to throttle back. Yet those systems we are installing have to be used by people, and in the end, it is the interaction between our agents and these machines that determines how valuable the application may be. So, yes, the technology has to meet a lot of technical specifications, but it also must fit in the culture of our organization. The ability for a system to match the skills of your agents may be the most important and least visible consideration. And it is a consideration not only when selecting the technology, but also when you decide how each system will be used.
Matching Staffing Profiles with Contact Center Systems
So where in a contact center technology plan do we have to pay close attention to the match between our staffing profiles and our systems? Almost any system used by your staff should garner this consideration (see the sidebar on page 17 for a discussion of IVR self- serve systems), but the major ones are knowledge management (KM), skills-based routing (SBR) and WFM.
Knowledge management is not as well defined as other applications. It can range from slick keyword searchable applications to sticky notes pasted to terminals, and everything else in between. In part due to the wide range of definitions, KM systems also deserve the most attention in this automation vs. expertise discussion. For the past decade or so, the prevailing wisdom has been that we need to move content into systems that are accessible by everyone. Rather than training agents to know everything, we simply train them to operate systems that know everything. At first glance, such a strategy certainly passes the logic test. Dig a bit below the surface, though, and you may uncover some unintended consequences caused by this shift of knowledge from people to systems. The central issue surrounding KM systems is that, no matter how good they may be, they are no match for a human brain. The best interaction still comes from the agent who knows all the content and procedures without having to access the system. It is more natural, conversational and efficient, and less prone to listening gaps. Of course, quite the opposite is true when the agent does not know the content. In this case, using a knowledge management system is far better than putting callers on hold, asking a neighbor or just “winging it.” The critical success factor, then, goes back to our central question: How automated do you want your call handling to be? You can build a KM system to be one-way and directive in nature (akin to a script), and insist that it be used on every contact. In a culture with a transient workforce and relatively simple calls, that may be appropriate. However, where agents are considered to be professionals, such a system does not match the culture, and call quality, efficiency and employee engagement can all suffer. More complex environments need KM systems that guide rather than direct, are collaborative in nature, and are used at the discretion of that highly trained agent.
Agents do not, of course, “use” SBR in handling contacts. It is a tool that determines what type of contacts they will receive, and the configuration possibilities are limitless. Here again, though, the skill level of the agents need to be considered in developing your strategy. Keeping SBR simple results in a limited amount of traffic being handled by a “non-primary” call handling group. That fits an environment that relies more on expertise than automation, since it limits the amount of content an agent must learn. Higher levels of SBR automation means that calls move around more, and agents are more often faced with a call that may not match their expertise. SBR usage, therefore, should mirror KM usage. Where you rely more on systems for content expertise, you are free to make greater use of SBR.
WFM systems can do some pretty amazing things, yet with all the sophistication there is little to no chance of any real success without a competent analyst at the helm. Yet many of these vendors try to sell their systems based on ease of use, touting such benefits as “one- click forecasting.” If you believe them, you may be inclined to staff your WFM roles with entry-level people… and you will never be pleased with the results. Regardless of the system in use, WFM roles require strong analytical skills and knowledge of advanced statistical principles—and people fitting this profile know better than to try to come up with a great forecast by simply clicking a box. WFM systems are a great example of the need to match systems with users. WFM systems that allow analysts to adjust forecasts and schedules, notate changes and apply different types of forecasting methods make the most of your investment in the right people. WFM systems that are not transparent and do not allow for creative input from the analysts will only serve to frustrate staff and decrease accuracy rates.
Choose Tools that Fit Your People
Today’s technology is better than ever and can pull off some amazing feats. Just because a system can do something, though, does not mean that it should. When contemplating new technology, it is easy to get lost in technical issues and vendor options. The first consideration, though, should be the culture of the operation. Tools should fit people, not the other way around. Team members involved in technology purchases need to understand the agent hiring profile just as much (or maybe even more) than they understand your technical environment. With that understanding, you have a better chance of getting not only the right systems, but the right mix of expertise and automation.
Since self-service via IVR does not “interact” with your contact center agents, it does not quite fit in the “automation vs. expertise” conversation as defined in this article. It does, however, fit quite well in the “how far do you go with automation” discussion.
Nearly everyone running a contact center today would like to see higher self-service utilization, and as a result, our IVR menus continue to grow. Yet reporting on IVR success is notoriously thin. Self-service success metrics may be provided on an overall basis, but few organizations report on success rates on an application basis. As a result, there are typically a few selections on a menu that drop a caller into self-service, yet rarely succeed in delivering resolution without help from an agent. That’s a classic example of over automation: forcing a customer to utilize self-service while delivering little if any value for the effort. To guard against that, contact centers should always have self-service reporting on an application basis.
An example is provided below:
If you look only at summary data, it would be easy to draw the conclusion that this is a successful self-service system,based on a 67% completion rate. Yet at a more detailed level, it becomes clear that the change address application is providing very little value. The low success rate is an almost certain sign that any savings being derived from it is more than offset by the dissatisfaction that 95% of customers in this application are feeling. Summary data does not help identify the problem—application level data is needed to determine how far to go with self-service automation.
– Reprinted with permission from Contact Center Pipeline, www.contactcenterpipeline.com
You must be logged in to post a comment.