A nice example of Swarming in support of a Service Desk
My research into the emergence of Swarming in enterprise technical support frequently brings up interesting examples of innovative new practices. Once such example came recently, in an interview with a technical support agent in South America.
The service desk in this example is chat-based: business consumers access it by initiating a text conversation. The conversation is initially automated on the service provider side: the consumer begins by talking with a bot. If the automated chat service cannot provide a solution, the system connects the consumer to a human support agent, who endeavours to resolve the issue.
Most service desks of this type achieve a pretty robust fix-rate. “First time resolution” is seen as an important benchmark. However, for a proportion of issues, immediate resolution will be beyond the skills or means of the frontline support agent. Usually, this results in the ticket being reassigned to a more specialised support team, who themselves will work to resolve or reassign the ticket. As a result, support teams are traditionally organised in a set of “tiers”.
As detailed in a longer article on this blog, advocates of Swarming argue that this tiered model is inefficient and slow. When the consumer’s ticket is reassigned, it usually arrives in a queue of work associated to the recipient team. Swarming, as a philosophy, holds that the tiers and queues should be replaced with a more dynamic, cross-functional structure, in which conversation between different areas of the support organisation replaces “over the fence” reassignments between siloed teams. There may be different types of Swarm in place in a single organisation, each adapting to different phases of the different lifecycle of different issues.
This particular service desk has implemented an interesting type of swarming, which I had not seen before. The front-line support agents are permitted to put the customer “on hold’ by asking them to wait for up to three minutes. While the conversation is paused from the point of view of the consumer, the support agent is able to join one of a set of open text chat “channels”.
Each channel is related to a particular technical subject area, and is crewed by one or more subject matter specialists. These experts are typically team members of specific product-line or technology support group, staffing the channel on rotation. They have both experience and deep knowledge of particular specialist topics. In those three minutes, the support agent can call on their knowledge. The result looks a little like this:
This neat methodology gives the support agent fast access to expertise that would typically only be deployed after the ticket related to the issue had dropped into the queue of a second-line or third-line support team (an action which ends the conversation with the consumer until a member of the receiving team gets round to processing that ticket). Often, those experts know the answer immediately: three minutes is absolutely sufficient. Swarming brings those three minutes right to the start of the interaction, rather than pushing them to the back of a queue.
The experience of this service desk closely echoes the benefits we have found from a similar, if not identical, process of frontline swarming. As a VP of Customer Support put it to me: “Swarming engages some of BMC’s best and brightest resources at the start of the problem solving process, versus at the end as in a traditional level 1/2/3 Support model. It has been transformational in improving resolution times”.
