Hiring · CX Leadership
How to hire for CX leadership, without screening for the wrong thing
Most CX leadership hiring screens for the same signal: polish. A strong CV, a consulting background, a framework ready before the question is finished. It looks like competence in an interview room. It often isn't what predicts whether the person can actually run your operation.
I have sat on both sides of this hire, as the candidate and as the person deciding, across 20 years and multiple European scale-ups. The pattern is consistent enough to be worth naming.
Polish is not the same as judgment
Someone with a strong consulting background can walk into an interview and produce a clean framework for any CX problem you describe. That is a real skill. It is also a skill that is fully visible in 45 minutes, which is exactly why it dominates interview scorecards: it is the easiest thing to evaluate, not the most important thing to have.
What is harder to see, and matters more, is judgment under a specific kind of pressure: a BPO missing SLA for the third month running, a board asking for an AI rollout in six weeks, a team that has quietly stopped believing their feedback changes anything. Frameworks do not tell you what someone does when the textbook answer and the organizationally possible answer are two different things.
What I actually screen for
Three things, in order.
- Operational scars, not case studies. Anyone can present a framework. Ask what they did when a vendor missed SLA for three straight months, or when the board cut the support budget mid-quarter. The answer to a real scar sounds different from the answer to a hypothetical: specific, a little uncomfortable, with a cost attached.
- Whether they have been on the floor. Not a requirement, but a strong signal. People who have taken calls, closed tickets, and been measured by AHT make different decisions about metrics later, because they know what a metric feels like from the other side. I started as an agent. It changed every decision I made as a director, particularly about which numbers to protect the team from and which to hold them to.
- What happens in month one. If the plan opens with a technology rollout, that is a warning sign. The operators who move first to diagnosis, before touching a single tool, are usually the ones who understand that most CX problems are structural before they are technical.
The interview question that filters fastest
Ask what they would do in their first thirty days, before they have seen your data. A candidate who answers with a plan is telling you they plan around assumptions. A candidate who answers by describing what they would go and look at first, escalation logs, repeat contact patterns, agent turnover by cohort, is telling you they diagnose before they prescribe. That difference alone eliminates most of the wrong hires I have seen made.
The best CX leaders I have hired or worked alongside share one habit: they ask what is actually happening before they say what should happen.
Why this matters more with AI in the picture
As AI absorbs the routine volume, the judgment calls left for a CX leader get harder, not easier: which escalations need a human, how to redesign a team's role when a third of their old workload disappears, how to tell a board that a deflection number looks good and the underlying experience does not. None of that is visible in a polished framework answer. It only shows up in how someone has actually handled pressure before.
I wrote more about the operational judgment this requires in Fractional CX Director: when you need one, what it costs, and when you don't, which covers the same evaluation logic from the hiring side of a fractional engagement.
What to do if you cannot tell
If an interview process leaves you unsure whether someone has real operational judgment or just a well-rehearsed framework, that uncertainty is itself useful information. It usually means the interview stayed hypothetical too long. Push one level deeper into a real, specific situation from their past, and watch whether the specificity holds or dissolves into generalities. Judgment shows up in details. Polish shows up in structure.