Easylab AI SARL

DPIA Screening Checklist

Based on EDPB Guidelines WP 248 rev.01 -- Data Protection Impact Assessment
Document ID: ELAB-DPIA-2026-001
Version: DRAFT v0.1
Date: March 2026
Author: [TO BE CONFIRMED]
Approved by: [TO BE CONFIRMED]

1. Purpose

This checklist determines whether a Data Protection Impact Assessment (DPIA) is required under Article 35 of the GDPR before launching a new project, product, feature, or service that involves the processing of personal data.

Under Article 35(1) GDPR, a DPIA is mandatory when processing is "likely to result in a high risk to the rights and freedoms of natural persons." The EDPB Working Party 29 Guidelines (WP 248 rev.01) identify nine criteria to assess this risk. As a general rule, if two or more criteria are met, a DPIA is required.

2. When to Use This Checklist

This checklist must be completed before any of the following:

The completed checklist must be retained as evidence of compliance, regardless of whether a full DPIA is subsequently required.

3. Project Identification

Project / Service Name
Project Owner
Date of Screening
Brief Description of Processing
Categories of Data Subjects
Categories of Personal Data
Legal Basis (Art. 6 GDPR)

4. WP 248 Screening Criteria

For each criterion below, check the box if it applies to the planned processing. Answer honestly -- underestimating risk exposes the company to regulatory sanctions.

1 Evaluation or Scoring

Does the processing involve evaluating or scoring individuals, including profiling and predicting?

Easylab examples:
  • EasyBlood: Health Score calculated from blood test results -- direct evaluation of an individual's health status
  • LinkedInScope: Analysis and scoring of LinkedIn profiles for recruitment suitability
  • Any AI-generated "quality score", "risk score", or "recommendation" based on personal data
2 Automated Decision-Making with Legal or Similarly Significant Effect

Does the processing involve automated decision-making that produces legal effects or similarly significant effects on individuals?

Easylab examples:
  • Automated rejection/acceptance of candidates based on AI-generated profile analysis
  • Health recommendations that could influence medical decisions (EasyBlood)
  • Any system output that directly determines access to a service, contract, or benefit without meaningful human review
3 Systematic Monitoring

Does the processing involve systematic monitoring of individuals, including observation of publicly accessible areas?

Easylab examples:
  • Session recording or heatmap tools on SaaS platforms (Hotjar, FullStory, etc.)
  • Continuous monitoring of user behavior/activity within a platform
  • Scraping or systematic collection of publicly available LinkedIn data
  • Audio recording of meetings (EasyBoard, EasyFund)
4 Sensitive Data or Data of a Highly Personal Nature

Does the processing involve special categories of data (Art. 9 GDPR), criminal conviction data (Art. 10), or other highly personal data (location, financial, communications)?

Easylab examples:
  • EasyBlood: Health data (blood test results, health scores) -- special category under Art. 9
  • EasyBoard / EasyFund: Audio recordings of meetings may contain opinions, political views, or other sensitive content
  • Any processing of biometric data, genetic data, racial/ethnic origin, trade union membership
  • Financial data discussed in board meetings (EasyBoard)
5 Data Processed on a Large Scale

Is the processing carried out on a large scale, considering the number of data subjects, volume of data, duration/permanence, and geographical extent?

Easylab examples:
  • A SaaS platform with thousands of end users across multiple organizations
  • Processing of all employees' data within a client organization
  • Linkeme platform if scaled to many users

Note: At current Easylab scale (early-stage, limited client base), most processing is unlikely to qualify as "large scale." This criterion should be re-evaluated as the business grows.

6 Matching or Combining Datasets

Does the processing involve matching or combining datasets from different sources in a way that exceeds the reasonable expectations of the data subject?

Easylab examples:
  • LinkedInScope: Combining publicly scraped LinkedIn data with client's internal HR data
  • Enriching user profiles by cross-referencing multiple data sources
  • Merging meeting transcripts with CRM data or other business systems
  • RAG (Retrieval-Augmented Generation) systems that combine personal data from multiple document sources
7 Data Concerning Vulnerable Data Subjects

Does the processing concern vulnerable individuals who may be unable to easily consent to or oppose the processing?

Easylab examples:
  • EasyBlood: Patients or individuals undergoing health assessments -- inherent power imbalance with healthcare providers
  • LinkedInScope: Job seekers or candidates in recruitment processes -- power imbalance with potential employers
  • Employees whose data is processed by their employer via Easylab tools
  • Children or elderly persons (if any Easylab service targets these groups)
8 Innovative Use of Technology or Organizational Solutions

Does the processing involve the use of new or innovative technologies, including AI, machine learning, or other emerging technologies?

Easylab examples:
  • All Easylab products: Use of generative AI (LLMs such as Claude, GPT) to process personal data
  • EasyBoard / EasyFund: AI-assisted transcription and summarization of meetings
  • EasyBlood: AI-driven health analysis from lab results
  • EasyClaw / OpenClaw: Autonomous AI agents with system access
  • RAG systems, vector databases (Qdrant), semantic search on personal data

Note: As of 2026, generative AI is still considered "innovative technology" by most DPAs. This criterion applies to virtually all Easylab AI-powered services.

9 Processing That Prevents Data Subjects from Exercising a Right or Using a Service or Contract

Does the processing prevent individuals from exercising a right, or using a service or a contract?

Easylab examples:
  • Credit scoring or eligibility checks that block access to financial services
  • Automated screening that prevents a candidate from progressing in a recruitment process
  • Mandatory data processing where refusal would deny access to an essential service

Note: This criterion is less likely to apply to current Easylab services, which are primarily B2B tools. However, it should be assessed if any service gates access based on personal data processing.

5. Screening Result

Total criteria checked_____ / 9

0-1 Criteria Met: DPIA Not Mandatory

A full DPIA is not legally required. However, conducting a DPIA is recommended if criterion 8 (innovative technology) is met, even alone. The EDPB emphasizes that the use of AI on personal data warrants careful assessment regardless of the formal threshold.

Retain this completed checklist as evidence of your assessment.

2 or More Criteria Met: DPIA IS MANDATORY

A Data Protection Impact Assessment must be completed before the processing begins (Art. 35(1) GDPR). Processing must not start until the DPIA is completed and, if necessary, mitigating measures have been implemented.

If the DPIA concludes that residual risk remains high despite mitigations, the supervisory authority (CNPD in Luxembourg) must be consulted under Article 36 GDPR before proceeding.

DecisionDPIA Required / Not Required (circle one)
Assessed by
Date
Reviewed by

6. Simplified DPIA Template

If the screening above indicates that a DPIA is required, use the following structure. For complex or high-risk processing, consider engaging external legal counsel or a qualified DPO.

Section A: Description of the Processing

Describe: nature, scope, context, and purposes of the processing. Include data flows, technologies used, data categories, data subjects, retention periods, and sub-processors involved.

Section B: Necessity and Proportionality

Assess: legal basis (Art. 6, Art. 9 if applicable), purpose limitation, data minimization, data quality, storage limitation, information to data subjects (Art. 13/14), right to object, data portability, lawfulness of international transfers.

Section C: Risks to the Rights and Freedoms of Data Subjects

Identify and rate risks by likelihood (unlikely / possible / likely) and severity (limited / significant / maximum). Consider: illegitimate access, unwanted modification, data loss, re-identification, discrimination, financial loss, reputational damage, loss of autonomy.

Section D: Measures to Address Risks

For each risk identified, document: the mitigating measure, its effect on likelihood/severity, the responsible person, and the implementation deadline. Include both technical measures (encryption, access controls, pseudonymization, logging) and organizational measures (policies, training, audits, incident response).

Section E: Conclusion and Approval

State whether residual risk is acceptable. If residual risk remains high, document the decision to consult the CNPD (Art. 36). Obtain sign-off from the project owner and, if appointed, the Data Protection Officer.

Luxembourg-specific: Check the CNPD's published list of processing operations requiring a DPIA (Art. 35(4) list). Some operations may be mandatory regardless of the WP 248 criteria count. See: cnpd.public.lu

7. Pre-Assessed Easylab Products

The following assessments are indicative and based on current product scope as of March 2026. They must be reviewed when product features change.

EasyBlood
#CriterionApplies?Rationale
1Evaluation or scoringYESHealth Score calculated from blood test results
2Automated decision-makingPossibleHealth recommendations could influence medical decisions; depends on implementation
3Systematic monitoringNoOne-time or periodic analysis, not continuous monitoring
4Sensitive dataYESHealth data (Art. 9 GDPR special category)
5Large scaleNoCurrently limited user base
6Matching datasetsNoSingle data source (lab results)
7Vulnerable data subjectsPossiblePatients may be considered vulnerable depending on context
8Innovative technologyYESGenerative AI (LLM) used for health data analysis
9Blocking a right/serviceNoNot applicable

Result: 3 criteria met (minimum) -- DPIA MANDATORY

LinkedInScope
#CriterionApplies?Rationale
1Evaluation or scoringYESAI-based profiling and scoring of LinkedIn profiles
2Automated decision-makingPossibleDepends on whether output is used for hiring decisions without human review
3Systematic monitoringPossibleSystematic collection from public profiles could qualify
4Sensitive dataPossibleLinkedIn profiles may reveal religion, ethnicity, political opinions indirectly
5Large scaleNoCurrently limited usage
6Matching datasetsPossibleIf combined with internal HR data
7Vulnerable data subjectsPossibleJob seekers are in a power imbalance with recruiters
8Innovative technologyYESGenerative AI used for profile analysis
9Blocking a right/servicePossibleIf used to filter out candidates from recruitment process

Result: 2 criteria met (minimum) -- DPIA MANDATORY

EasyBoard / EasyFund
#CriterionApplies?Rationale
1Evaluation or scoringNoTranscription and summarization, not scoring individuals
2Automated decision-makingNoOutput is informational (minutes), not decision-making
3Systematic monitoringPossibleAudio recording of meetings could qualify depending on scope
4Sensitive dataYESMeeting audio may contain sensitive opinions, financial data, strategic discussions
5Large scaleNoCurrently limited client base
6Matching datasetsNoSingle data source per processing operation
7Vulnerable data subjectsNoBoard members and executives are not typically considered vulnerable
8Innovative technologyYESGenerative AI (LLM) + speech-to-text for meeting transcription
9Blocking a right/serviceNoNot applicable

Result: 2 criteria met (minimum) -- DPIA MANDATORY

Easylab.ai Website (Corporate Site)
#CriterionApplies?Rationale
1Evaluation or scoringNoStatic website, no profiling
2Automated decision-makingNoNo automated decisions
3Systematic monitoringNoStandard analytics only (if any)
4Sensitive dataNoContact form data only
5Large scaleNoSmall website
6Matching datasetsNoNo dataset combination
7Vulnerable data subjectsNoGeneral public, B2B visitors
8Innovative technologyNoStandard website technology
9Blocking a right/serviceNoNot applicable

Result: 0 criteria met -- DPIA not mandatory

8. References

DRAFT DOCUMENT -- v0.1 -- March 2026
This document is a working draft pending legal review. It should not be relied upon as a definitive compliance document.
Easylab AI SARL | 55, allée de la Poudrerie, L-1899 Roeser, Luxembourg