Where the information came from.
Method & sources
Evidence first.
Interpretation second
ResumeAI Lab separates source evidence, structured observations and our own interpretation so you can see what a claim is based on and where its limits are.
What the source actually supports.
What ResumeAI Lab concludes, with our reasoning connected to evidence.
How evidence becomes a published claim
Each claim follows the same structured process, from source to published guidance.
- 1Source
Official statistics, employer material, public guidance or other reviewed evidence.
- 2Capture
Record the relevant fact or observation together with its origin and date.
- 3Classify
Separate source type, requirement strength, evidence category and provenance.
- 4Interpret
Identify what the evidence reasonably supports and what it does not.
- 5Publish
Write the user-facing claim with the right scope, attribution and limits.
Different evidence has different jobs
ResumeAI Lab brings several evidence types together. Each type can tell us different things and has limits.
Official statistics
- Example
- Statistics Denmark
- Purpose
- Provide wider labour-market context and comparable official indicators.
- Supports
- Employment, unemployment and vacancy movement in the published series.
- Does not support
- They do not predict an individual's hiring outcome.
Employer & vacancy evidence
- Example
- Reviewed public vacancies and employer material
- Purpose
- Show what employers in the focused sample are asking for.
- Supports
- Observed roles, locations, career stages and application instructions.
- Does not support
- The sample is not every Danish employer and does not create a national ranking.
Public application guidance
- Example
- Workindenmark, universities and public career services
- Purpose
- Document application conventions and recommendations.
- Supports
- Source-specific guidance and recurring, attributable positions.
- Does not support
- One recommendation does not become a universal Danish rule.
ResumeAI Lab synthesis
- Example
- Our labelled interpretation
- Purpose
- Combine evidence into practical explanations and choices.
- Supports
- A bounded conclusion that stays connected to its sources.
- Does not support
- It is not presented as a direct external-source finding.
See the distinction in practice
The same topic can involve different source positions. We show what those sources support and what we conclude.
Source position 1
Professional-picture guidance
Workindenmark advises applicants to include a professional picture when writing a CV for a Danish employer.
View Workindenmark guidanceSource position 2
Photo treated as optional
International House Odense lists a photo as optional in its Denmark job-application guide.
View International House Odense guidanceWhat the sources support
Danish guidance does not express one universal rule
- Photo guidance differs by source and context.
- Employer instructions still matter.
- A single universal rule is not supported.
ResumeAI Lab synthesis
Photos are common in Danish CV practice, but they are not universally compulsory
Follow the employer's instructions first. If there is no instruction, make a deliberate, professional choice.
How reviewed vacancies become market evidence
We use a governed process so observations stay consistent, traceable and current.
- 1Collect
Review permitted public vacancy and employer evidence.
- 2Structure
Record governed fields with a shared taxonomy.
- 3Keep provenance
Retain source, employer, date and raw provenance.
- 4Refresh
Separate current vacancies from closed or historical observations.
We track not only the source, but how it may be used
Public accessibility is not automatically treated as permission for unrestricted reuse.
- Source
Where the information comes from
- Permitted use
How the source may be used or cited
- Structured observation
What we record and retain with provenance
- Public output
What we publish, with clear scope and limits
Evidence also tells us when to stop
- A focused vacancy sample is not the whole labour market
- A CV convention is not necessarily a requirement
- An illustrative CV is not evidence that someone was hired
- Missing candidate evidence is not filled with invented achievements
- Observed patterns do not become hiring probabilities
- Older observations are not presented as currently open vacancies
Evidence changes, so we check it again
Different evidence types are refreshed in different ways. We do not claim daily or real-time checking.
Last methodology review: 13 September 2026
Vacancy evidence
Reviewed prospectively through the recurring dataset process. Active and historical observations stay separate.
Official indicators
Checked for newer official releases during the recurring market-update cycle.
Guidance & source-linked claims
Rechecked when pages are materially updated and when continued source support matters.
Key source families
These source families shape the evidence we use and the guidance we publish.