Staffing Support in the Age of AI
Many thanks to Kt Doan for suggesting this topic. (If you’d like to suggest a topic, just email me your suggestion).
We are seeing AI tools revolutionizing self-service, whether through chatbots on support websites or, more and more, commercial LLMs. And internally, AI tools are allowing agents to find answers much more easily throughout external, customer-facing, and internal sources. What does that mean for staffing?
Fewer but more sophisticated level 1 agents
Much of the discussion (and hand-wringing) has been around the disappearance of the traditional entry-level jobs. This may not be a bad thing, since these jobs are mind-numbing, and usually managed in spirit-crushing ways through purely quantitative metrics that have little to do with customer satisfaction, and even less with employee satisfaction. Let’s start there: it’s true that team members who today handle repetitive how-to questions or automatable requests will no longer be needed. But we will continue to need agents who can assist customers in finding answers and overcoming automation obstacles. And these agents will contribute to knowledge management, likely not by directly writing knowledge articles (how old fashioned!) but by allowing AI agents to capture their experiences.
I believe that most agents in tier-1 jobs have the capacity and desire to do more complex work, so there’s hope and growth for them. But overall, support organizations will need fewer of them.
A stronger push towards tierless support
Many support organizations have already disbanded tiering of agents. For the rest, as the AI revolution means that we will have significantly fewer tier-1 agents, the very reasoning behind tiered support will vanish. This is a positive development for customers, who hate tiering, and it for agents, who can more easily explore career growth.
More meaningful metrics
Currently, tier-1 agents are often measured through metrics that don’t connect to customer satisfaction at all, like AHT (average handle time), and others with only a vague connection to it (I’m looking at you, manual “call quality”). As more and more work becomes automated, and as customer sentiment can be measured in real time, I predict (hope!) that we will move towards more meaningful metrics centered around true customer success and productivity.
Continued hiring of junior staff
A big worry of the AI revolution is that early-career individuals won’t be able to find jobs. I’m a little perplexed by that. Surely we’ve all noticed that young people are very much at ease with AI technology and adept at learning more. They will be hired into different jobs than we have been used to, but they will be hired, especially if they have been smart enough and lucky enough to pick up some technical skills.
What does that mean today?
The AI revolution is swift but not absolutely abrupt. We have time to adapt. Here are five key tasks for right now.
- Continue to automate with AI wherever it makes sense. If you need inspiration, look here.
- Adapt your metrics. There’s no reason for dumb metrics, even if the tier-1 jobs are still quite menial for now.
- Identify tier-1 agents who can do more complex tasks. You may be surprised by how many can.
- Retool your staffing model. Incoming volume will be higher; productivity (if measured by interactions per agent) will be lower; tiering may well disappear.
- Up-train. If your tier-1 agents act more like escalation leads, they will need stronger soft skills.
How is your staffing model changing as AI automation increases?
