Consulting and recruiting
IT staffing firms: technical assessments and interview data
By SourceX Editorial · Reviewed by Noah Loul ·
Short answer
An IT staffing firm's data includes skills-based job orders, technical screens, coding assessment results, interview feedback and placement outcomes. The privacy limit is sharp: scores, code submissions and recorded interviews are personal data tied to named candidates and are usually excluded. Firm-written question banks, rubrics and de-identified outcome codes are what typically remain licensable.
Key takeaways
- Technical assessment results are personal data about the candidate, even when they look like plain scores.
- A candidate's code submission was written by the candidate, so the firm may not own it.
- Recorded video interviews raise extra concerns, including biometric privacy laws where face or voice is analyzed.
- Question banks, rubrics and skills taxonomies written by the firm's own staff are usually the firm's to license.
- Assessment vendor terms decide what you may do with results stored on their platforms.
What records does an IT staffing firm hold?#
An IT staffing firm holds a denser technical record than most recruiters: job orders written in skills and tools, technical screens run by recruiters or engineers, coding assessments, client interview outcomes and contractor placements that extend or convert. Those records are spread across the ATS, an assessment platform, video interview tools and the back office.
Each family carries a different mix of firm knowledge and candidate data, and that mix decides its treatment before anyone talks about value.
| Record family | Typical system | Personal data? | Usual treatment |
|---|---|---|---|
| Skills-based job orders | ATS | Client contacts only | Generalize client identity |
| Technical screen scorecards | ATS forms | Yes | Exclude or reduce to outcome codes |
| Coding assessment results and submissions | Third-party assessment platform | Yes | Exclude |
| Recorded video interviews | Video interview tool | Yes, including face and voice | Exclude |
| Question banks and rubrics | Shared drive, assessment platform | No, if written by staff | Often in scope |
| Client interview outcomes | ATS | Yes, linked to the candidate | De-identified codes only |
| Contractor placements and extensions | ATS, back office | Yes | De-identified events only |
Why technical assessment results count as personal data#
Technical assessment results count as personal data because they describe an identifiable person's abilities. A candidate's result on a data structures test is as personal as a performance review, and it can shape that person's career.
Removing the name does not always help. Assessment records usually carry timestamps, the role applied for, the client and sometimes the candidate's code, which can include a distinctive style, comments or references to past employers. Combined, those details can lead back to one person.
Many firms also store results on the assessment vendor's platform rather than in the ATS. The vendor's terms then govern export, retention and reuse, and some limit results to the hiring process. Read those terms before assuming results are yours to license.
- 01Your work records
- 02Personal details removed
- 03You approve the release
- 04Licensed to AI developers
Who owns a candidate's code submission?#
A candidate's code submission is usually authored by the candidate, not by the firm. Copyright treats a work as made for hire when an employee creates it within the scope of employment, or when a commissioned work in a listed category is covered by a signed written agreement. A job candidate is usually neither your employee nor party to such an agreement.
Some assessment platforms address ownership in their candidate-facing terms, and some reserve rights for themselves. Either way, ownership is uncertain enough that candidate code is normally excluded from a license, separate from the privacy question.
Where privacy law draws extra lines#
Privacy law draws extra lines around interview recordings and anything tied to a candidate's identity. Video interviews capture face and voice, and where a tool analyzes those features, biometric privacy laws may apply. Illinois's Biometric Information Privacy Act, for example, may apply where face or voice geometry is captured; it provides liquidated damages per violation, and the Illinois Supreme Court has held that a person need not show harm beyond the violation itself to sue.
Applicant data more broadly has been within California's CCPA for covered businesses since the employee and applicant exemption expired on January 1, 2023, while some other states exclude employment-context data. Which laws apply depends on where candidates are and what the records hold, and counsel assesses that deal by deal.
- Recorded video and audio interviews, and their transcripts.
- Candidate code submissions and take-home projects.
- Individual scores, rankings and pass or fail flags tied to a person.
- Proctoring data such as webcam snapshots or browser activity logs.
- Recruiter and engineer comments on a candidate's performance.
What an IT staffing firm can usually license#
What an IT staffing firm can usually license is material its own staff created about evaluating technical skill, plus de-identified records of how the hiring process ran. That material reflects the firm's domain expertise rather than any candidate's personal information.
The strongest version links the two: a question or rubric, the skill family it tests, and de-identified outcomes showing which skill families advanced to client interviews and starts. That pairing shows how assessment connected to hiring decisions without showing who was assessed.
| Record | Why it can qualify | Condition |
|---|---|---|
| Firm-written question banks | Created by staff within their roles | Remove questions copied from vendors or public sources |
| Scoring rubrics and interview guides | Show how skills are judged | Remove client-specific content |
| Skills taxonomy and job order requirements | Show how technical demand shifts | Generalize client identity |
| De-identified pipeline events | Show stage timing and outcomes | Candidate IDs replaced, rare details generalized |
| Outcome codes by skill family | Link screening to client decisions | Codes only, no individual scores |
How to sort assessment data before any review#
Sorting assessment data before any review means finding every place results live and labeling each source as firm-authored or candidate-generated. IT staffing firms often discover assessment material in more systems than expected, because engineers who run technical screens keep their own scorecards and question lists.
The output is a short inventory that counsel can sign off on, with each source marked in scope, out of scope or needing review.
- List every assessment, video interview and scheduling tool in use now or in past years, with its contract.
- Find where engineers keep screening questions, scorecards and debrief notes outside the ATS.
- Separate firm-authored content, such as questions and rubrics, from candidate-generated content, such as answers and code.
- Record what each vendor's terms say about export, retention and use of results.
- Confirm which fields link assessments to job orders, so outcomes can be kept as codes.
- Note any client that supplied its own test questions, since those usually belong to the client.
Illustrative: a cloud and data engineering staffing firm#
Illustrative: a fictional IT staffing firm places cloud, data and security engineers on contract with enterprise clients. It runs coding assessments through a third-party platform, records first-round video screens, and employs senior engineers who wrote its technical question bank and scoring rubrics.
When the owner considers a data license, counsel excludes every video recording, every candidate code submission and all individual scores. The assessment vendor's terms confirm that results are available to the firm for hiring use only. The proposed package contains the question bank and rubrics, a skills taxonomy built from years of job orders, and de-identified pipeline events showing which skill families reached client interviews and starts.
How SourceX approaches IT recruiting records#
SourceX approaches IT recruiting records by separating firm-created material from candidate material in the Rights step of the SourceX five-step transaction. Assessments, recordings and candidate code are treated as excluded unless counsel finds otherwise, and the review moves on to what the firm itself authored.
Under the SourceX Enterprise Data Value Framework, a question bank written by working engineers can rate well on domain expertise and human-generated signal, while privacy burden counts against anything tied to individuals. The SourceX Evidence Packet records which record families were excluded and why.
Frequently asked questions
Can we license our technical interview question bank?
Often, if your staff wrote it within their jobs and it does not copy vendor or published material. Check for questions borrowed from assessment platforms, books or websites, and remove anything client-specific. Questions in wide circulation may add less, but rubrics and model answers add context.
Are third-party assessment results ours?
They are usually available to you for hiring decisions, but the platform's terms decide export and reuse. Some terms limit results to the hiring process or keep rights in the test content with the vendor. Read the current agreement and the candidate-facing terms before treating results as firm records.
Does code our contractors write at client sites count as our data?
Usually not. Contractors placed with clients work under the client's direction, and staffing agreements commonly assign their work product to the client. That code belongs in the client's systems, not in a staffing firm's data package.
Should we change candidate consent language going forward?
Many firms review it. Clear notices about how assessment data is stored, kept and used help with compliance whatever you decide about licensing. FTC staff have warned that quietly changing terms to allow AI training on data already collected may be unfair or deceptive, so draft changes with counsel.
Do AI developers want coding interview data?
Some look for material on how technical skill is assessed, such as questions, rubrics and anonymized pipeline outcomes. Real candidates' code and recordings are a different matter, with ownership and privacy problems most firms cannot resolve. Start with the content your firm authored.
Sources
- A work made for hire arises when an employee creates the work within regular duties, or when a work in certain statutory categories is specially commissioned under an express written agreement. Source
- The Illinois Biometric Information Privacy Act lets a prevailing party recover liquidated damages per negligent or intentional violation, or actual damages if greater. Source
- In Rosenbach v. Six Flags (2019), the Illinois Supreme Court held that a person need not allege actual injury beyond a BIPA violation to be an aggrieved party entitled to sue. Source
- The CCPA employee and business-to-business exemptions expired on January 1, 2023. Source
- FTC staff warned in February 2024 that adopting more permissive data practices such as AI training through a surreptitious, retroactive change to terms or privacy policy may be unfair or deceptive. Source
Related resources
- QuestionDo AI labs buy financial data?
- InsightAP automation and procurement software vendors: licensing invoice data
- IndustryBPO & contact centers data
- QuestionCan anonymized data be licensed?
- InsightCan financial advisors sell their data to AI companies?
- InsightCan data licensing affect my cyber insurance?
See if your company qualifies
A short company assessment. No data uploads are needed.