The assessment
We assess people against the work, and write down why
The same work sits behind both of our services. A client does not receive a profile. They receive reasoning they can disagree with.
The process
Four steps from the requirements to the note
We start with the work to be done, not with a list of names.
- 01
The requirements
We go through the engagement with you and turn it into requirements that can actually be assessed. Vague wishes either become concrete or come out.
- 02
The search
We look through professional networks and public profiles across Nordic and European fields, including people who are not looking for work.
- 03
The assessment
Every relevant person is assessed against the same requirements. We record what is met, what is partly met, and what we are unsure about.
- 04
The note
You get one assessment note per name. One of us reads it before it is sent. The decision is always yours.
The document
This is what an assessment note looks like
The format is the same whether the engagement is a hire or evaluation work for a lab.
- Role
- Senior backend engineer, payments infrastructure
- Field
- Distributed systems, payments
- Reference
- TAL-EX-001
Requirements taken from the role
- Has owned a payments system in productionMet
- Has built idempotent handling of repeated callsMet
- Works in Norwegian or EnglishMet
- Experience with Norwegian financial regulationPartial
- Available within three monthsMet
Reasoning
Seven years on payments infrastructure, the last three owning reconciliation and recovery after failure. Wrote the retry logic behind it and can explain why it looks the way it does. The regulatory side is thinner: they have worked against those requirements without owning them. We think that is learnable in this role, and we flag it rather than hide it.
The example is constructed to show the format. It describes no real person, and none of the values are measurements.
The chain
Three layers before anything reaches you
There are two of us. That keeps the chain short, and it makes it possible to say exactly what each layer does and does not do.
- 01
Search and sort
- Does
- Software searches professional networks and public profiles and sorts the hits against the requirements we wrote down.
- Does not
- Decides nothing. The sort sets the order we read in, not who ends up on the list.
- 02
Written assessment
- Does
- One of us works through every relevant person against the same set of requirements and writes down what is met, what is partly met, and what we are unsure about.
- Does not
- Produces no score, and hides no weakness. If something is thin, it goes in the note.
- 03
Read before it ships
- Does
- The other one reads the list as a whole before it goes out, checking that the requirements were applied the same way to everyone.
- Does not
- Decides who you talk to. That is yours.
Criteria can change
Change the requirements and we assess again
Because the requirements are written down, they can change mid engagement without the work starting over.
- 01
You read the reasoning
You see why each person was assessed the way they were, and against which requirements.
- 02
You adjust the requirements
If one is stricter than you meant, or something is missing that only became clear once you saw real candidates, you tell us.
- 03
We assess against the new set
We go back through the same people with the updated requirements, and tell you what changed.
This is the practical difference between an assessment and a profile. A profile you can only accept or reject. An assessment you can argue with, and then run again.
Rules we hold to
What never appears in a note
- No score
- We do not rank people on a scale we cannot defend. A requirement is met, partly met or not met, and the reasoning sits next to it.
- No hidden weaknesses
- If something is thin, it goes in the note. A shortlist containing only strengths is a sales brochure, not an assessment.
- No automated decision
- Software searches and sorts. A person reads the list before it is sent, and the client decides who goes forward.
- No data we do not need
- We process what is professionally relevant to the engagement and no more. The privacy policy sets out exactly what that is.
Human oversight
Why a person always reads the list
AI used in recruitment and selection is high risk under Annex III of the EU AI Act. We would take this view regardless: a system that sorts people should be able to explain itself, and somebody should be accountable for what gets sent on.
In practice the software does the search and the groundwork, while the assessment is written down and read by one of us before it leaves the building. We make no solely automated decisions producing legal effects for the individual.
We do not sell compliance as a feature, and we are not going to claim we are finished with a regime that is still being phased in. We describe what we actually do, and follow what applies.
Contact
Tell us what you need
We answer every enquiry. If we are not the right choice for you, we will say so.