Contact
A retail chain opening for the season might receive several thousand applications across hundreds of near-identical roles. The bottleneck is the workload, the manual review of every applicant who each deserves a fair shot. We still hear of the store managers who look at the first ten applicants, pick five and scrap the rest if they find a hire that will suffice. The risk here is not only a bad hire and negative employer brand experience; it is a good candidate lost to a slow, inconsistent, or arbitrary screen before anyone ever spoke to them.
A private office hiring a single executive assistant to a principal has the opposite problem. There may be only a handful of genuinely suitable people, the brief is confidential, discretion and chemistry matter as much as competence, and the "assessment" is really a long exercise in trust and judgment. Throughput is irrelevant. What matters is depth, discretion, and a human who knows the market personally and builds the rapport with the manager.
Both hiring - but with two completely different problems to solve. Hiring isn't a one-size-fits-all and AI in hiring shouldn't be treated that way either.
When applications arrive in the hundreds or thousands, structured, skills-based interviewing is the method that holds up. This is what Hubert is built for. Every candidate completes the same competency-based interview, and every answer is assessed by deterministic models: the same candidate answer produces the same score, every time, with every score tied back to a specific response for a full audit trail. That consistency is what makes a shortlist fair, and what makes it defensible when a stakeholder or a regulator asks how you decided.
The outcomes at scale are concrete. ManpowerGroup reduced screening time for recruiters by 67%. OKQ8 reached an average time-to-shortlist of 57 minutes, with 90% of candidates completing the interview and a 9/10 candidate satisfaction score. NSS Group increased hires by 50% among candidates who would never have passed traditional CV screening, because a structured interview surfaces what people can do rather than filtering on what their CV happens to say. Across deployments, candidates rate the experience 9/10, and the recruiter always makes the final decision.
This is the end of the spectrum where a machine genuinely does the job better than a stretched human team can: not because it is smarter, but because it is consistent across thousands of people in a way humans cannot be.
Now take the executive assistant hire, or a confidential senior appointment, or a specialist role where the qualified pool is tiny. Here, structured screening at scale adds almost nothing, because there is no scale. What the hire needs is a person who knows the market, can approach passive candidates discreetly, reads chemistry and fit in the room, and manages a delicate, human process end to end.
That is the world specialist search firms work in. Auria Heanley, Co-founder of Oriel Partners, a London-based PA & EA recruitment agency, built her practice around exactly this kind of hire, placing personal assistants, executive assistants, and family-office staff into households and principals' offices where confidentiality and personal fit matter more than throughput. Her team's approach to private PA recruitment is built on precisely the long-term market relationships and discretion that no volume-screening tool is designed to replace. For those roles, that is exactly the right tool, and no interview automation should pretend otherwise.
We say this plainly because it is true, and because pretending a single method fits every hire is how talent teams end up disappointed with tools that were never built for their problem.
Point volume automation at a one-of-a-kind senior hire and you get a process that feels impersonal for a candidate who expected to be courted, and an assessment that measures the wrong things. The role calls for judgment and discretion; don't bring it to a funnel motion.
Try a manual, human process at five thousand applicants and something quietly worse happens. Under that much pressure, screening becomes inconsistent by necessity: the hundredth CV of the day is not read like the first, shortcuts creep in, and bias fills the gaps where time runs out. Good candidates are lost not through a considered decision but through sheer volume nad recruiters are felt burnt out and demotivated. That is the exact failure structured interviewing is designed to remove.
A framework isn’t needed but here’s a reminder of some simple questions that should be obvious enough:
Most of the time the answer is obvious.
The same organization screening thousands of frontline applicants also hires senior leaders and confidential support roles where a specialist search partner is the right call. Maturity in talent acquisition is not picking a side in the AI debate; it is knowing which method each role deserves and refusing to force one tool onto every problem.
Hubert sits firmly at the high-volume, structured end, and we are clear about where that ends. Our job is to make the initial screening stage fair, fast, and defensible when the numbers are large, and to hand recruiters shortlists they can stand behind. The final decision always stays with your team, and when a role belongs to the human-search end of the spectrum, the honest answer is to treat it that way.
Get the match right and the tension disappears. Structured interviewing does what it is built for, human expertise does what only it can, and every candidate, whether one of five thousand or one of five, gets a fair shot at the role.