For decades, we have compared CVs with job advertisements. If this method keeps producing poor matches, why are we using AI mainly to speed it up?
People write in their CVs what they have done before. Companies write in job ads what a new employee is expected to do. Then candidates consider whether the job might appeal to them, and employers consider whether the candidate might be able to do it.
It sounds reasonable. But neither document describes the real situation.
A CV usually does not show what kind of work gives someone energy, which tasks they do well but with considerable inner effort, or which capabilities have gone unused in their previous roles. A job ad rarely explains how the work is actually divided, who the person will work with every day, which decisions they can make themselves, or what problems already exist in the team.
Both sides try to make a good impression. Sometimes they knowingly present things in a better light. Often, candidates and employers sincerely believe what they say. They simply do not yet know what they should have asked. The result can be an honest conversation based on two incomplete descriptions.
And then we call the result an assessment of fit.
Bad statistics do not explain the whole cause. But they are no reason to keep doing things the same way.
A 2017 US CareerBuilder survey found that 74% of surveyed employers said they had hired the wrong person for a position. Coverage of an Estonian labour market survey in 2023 reported that 73% of employees were open to job offers. In that coverage, the 87% figure applied only to respondents who rated their financial situation as poor. These figures do not mean that 74% of employers are currently dissatisfied with all their employees or that 87% of all workers want to change jobs. They also do not prove that CV-based recruitment caused every poor match. But they give us reason to ask why bad hiring decisions and readiness to leave are such familiar problems.
What do we do with this knowledge? We write better job ads. We teach candidates to write better CVs. We screen applications faster. With AI, we can do all of this at a much greater scale than before.
But if we are still comparing a short summary of a person’s past with a short summary of a job’s future, we have accelerated an old method. A faster decision does not turn incomplete starting information into an accurate decision.
How quickly we reach an outcome that neither side is happy with cannot be the main measure of recruitment success.
A person is not a job title
A job is not just a block of text in an advertisement either. It consists of tasks, responsibility, workload, decision-making authority, people and environment. These change over time. People change too, learn, and may discover capabilities that their previous work never brought out.
That is why it is not enough to place someone into one of four or five types. Such a model can give us language for a conversation, but it cannot by itself tell us how a particular person will handle particular tasks in this particular team. Gut feeling alone is not enough either. A manager’s or recruiter’s observations are valuable, but they need to be checked against the actual requirements of the work and information from several sources.
Accuracy requires both sides to see the fullest possible picture honestly. A company needs to know what work must be done, not only which position is vacant. A person needs to be able to describe what they can do, what they want to do, the environment in which they work well, and what they could still learn. Uncertainty remains, but the decision does not have to be based only on two pieces of sales copy.
Sometimes the right candidate is already in the company
Imagine that a company is looking for someone new for position A. While assessing candidates, it turns out that one of them would fit position B better, even though it is already filled. At the same time, the person in position B has long been ready to take on the tasks of position A and could use their capabilities much better there.
If we look only at the vacancy and the CVs received, we miss this possibility. When we see people’s capabilities, the actual tasks and the team’s needs together, a different decision becomes possible: one person moves forward, another finds a better-fitting role, and the company does not waste the talent it already has.
This is not about moving people around arbitrarily. Such a change requires a conversation, the person’s own willingness, and a check of their skills against the work requirements. But we cannot even discuss the possibility if we do not know what people can and want to do.
Sparkly’s central question
Sparkly starts with this question: how can we make fewer wrong decisions about people instead of making the same decisions faster?
To do that, we need to look at the person, their capabilities and development potential, the tasks and responsibilities, and the team composition together. This is technically complex because no simple test or type table can describe all the variation. Sparkly synthesizes different analytical perspectives to give managers, HR leaders and recruiters more information before an interview and better questions to ask during it. The results are decision-support tools: they need to be checked through conversation and the actual work.
This also means being honest with candidates. Every person can be valuable, but not equally suited to every job and every combination of people. “Not suited to this position” does not mean “a bad person.” Sometimes it means they could thrive somewhere else. The same applies to a company’s current employees.
The cost of a bad hire can be high: lost time, recruiting again, training, extra workload for others, mistakes and lost revenue. An annual cost of €50,000–€150,000 for one case should not be presented as a general fact; it needs to be calculated for the company and role. That makes it all the more important to understand what kind of fit we check before deciding.
We do not need to hire as many people as possible, as quickly as possible. We need to understand as accurately as possible what work needs doing, who could do it well and in what team they can realize their capabilities.
Maybe it is time to stop rushing faster in the wrong direction. Let’s slow down for a moment, look at the real work and real people, and make a decision that benefits both sides.
If you want to see which capabilities are unused in your current team or how to compare candidates with the actual work, contact me. Let’s talk about your company’s specific situation.