Using AI-powered software? As the business deploying these tools, you are the deployer under the EU AI Act — and deployer obligations are your responsibility, not the vendor's. These guides explain exactly what you need to do for each tool.
Take the 8-question HR AI Compliance Checker — identify your gaps and get a personalised action plan in 2 minutes.
Start free assessment →Enter your revenue and AI use cases — calculate your exact EU AI Act fine cap under Articles 5, 6–49, and 50 in 60 seconds.
Calculate fine exposure →Article 50 transparency disclosure required by 2 Aug 2026. Ready-to-use disclosure templates included.
Read guide →Answer Bot and messaging bots need Art.50 disclosure. Escalates to high-risk if used for credit, benefits, or HR decisions.
Read guide →B2B chatbot playbooks require Article 50 disclosure before first message. Steps and ready-to-paste templates for Drift playbooks.
Read guide →E-commerce chatbot guide. Shopify and WooCommerce store owners — what to add to Lyro in 10 minutes.
Read guide →Freshworks customers guide. Covers Freshchat, Freshdesk, and Freshservice — which AI features trigger Article 50.
Read guide →Two obligations: chatbot disclosure (Art.50(1)) and AI content labelling (Art.50(2)). Marketing and sales teams guide.
Read guide →Crisp chatbot scenarios and MagicReply AI guide. What triggers Article 50 and what doesn't.
Read guide →LiveChat paired with ChatBot product. Which AI interactions trigger Article 50 and steps to add the disclosure.
Read guide →Instagram DM, WhatsApp, and Messenger automation. Article 50 applies to all channels — one disclosure per flow entry point.
Read guide →Facebook and Instagram chatbot automation. Article 50 disclosure required at start of each automated conversation flow.
Read guide →Free live chat with chatbot automation. If any automated triggers are active, Article 50 disclosure applies. 10-minute fix.
Read guide →Internal use vs. customer chatbot vs. AI content publishing — three different obligation levels explained.
Read guide →Standard dev use is minimal risk — no Article 50, no high-risk. What engineering teams actually need to do.
Read guide →Internal knowledge work is minimal risk. When published AI-generated content triggers Art.50(2) labelling.
Read guide →Writing assistance is minimal risk. One scenario where Art.50(2) kicks in. Data protection considerations for business accounts.
Read guide →Call recording + AI scoring of sales reps is Annex III Category 4 when used for performance decisions. What sales leaders must do.
Read guide →Internal Workspace use = minimal. Customer chatbot via API = Article 50(1). AI content at scale = Article 50(2). Three scenarios explained.
Read guide →AI content generation for marketing. When does published Jasper content need an Article 50(2) label? The human-review threshold explained.
Read guide →Meeting transcription AI. GDPR is the bigger concern. When Otter AI Meeting Agents join calls, Article 50 applies.
Read guide →AI Meeting Summary = minimal. AI Chat responses = Article 50(1) if customer-facing. Feature-by-feature breakdown.
Read guide →Sales outreach AI = limited risk. Conversations rep scoring used for employment decisions = potentially high-risk Annex III.
Read guide →Contract analysis AI is minimal risk. AI-generated employment or credit agreements without legal review escalates risk. Guide for legal teams.
Read guide →SMB CRM AI guide. Sales Assistant and deal scoring are minimal risk. Only chatbot feature triggers Article 50 disclosure.
Read guide →Meeting transcription AI. GDPR-first obligations: inform participants before recording. Analytics used for employee evaluation escalates risk.
Read guide →Async video with AI transcription and summaries. Minimal risk for internal use. Article 50(2) applies only to AI-generated video content published externally.
Read guide →AI-generated images and creative content require Article 50(2) labelling. Deepfakes of real people need explicit disclosure. Content Credentials help satisfy both.
Read guide →Finance and ESG reporting AI. Regulated outputs require qualified human sign-off. AI-assisted financial disclosures need governance regardless of formal risk level.
Read guide →Legal AI contract review. Minimal EU AI Act risk for most use. Professional responsibility requires lawyer review of all AI-extracted provisions.
Read guide →AI sales sequences = limited risk. Kaia call scoring used for rep management = potentially high-risk Annex III. Audit which features you use.
Read guide →Zia AI predictions are internal analytics — minimal risk. SalesIQ chatbot requires Article 50 disclosure. SMB-focused guide for European Zoho users.
Read guide →B2B prospecting with EU contacts creates GDPR obligations first. AI sequences at scale create Article 50(2) risk. GDPR legitimate interests assessment required.
Read guide →AI video interview analysis triggers Annex III Category 4. Candidate disclosure, human oversight, and log retention required.
Read guide →Workday Recruiting and Skills Cloud are high-risk. HCM payroll and scheduling are not. Know which features apply.
Read guide →AI Scoring and Interview Intelligence are high-risk. Job posting and scheduling automation are not.
Read guide →AI Recommendations, Spotlights, and skills matching are high-risk. The upstream AI challenge explained.
Read guide →US, UK, or Indian company? 5 questions to determine whether the EU AI Act applies to you under Article 2(1)(c).
Check scope →Recruiting AI, Opportunity Marketplace, and Workforce Analytics are high-risk. Enterprise deployer obligations explained.
Read guide →SMB-focused guide. Smart Recommendations and candidate scoring are high-risk. 5 steps for growing companies.
Read guide →Emotion inference in the workplace prohibited since Feb 2025. Urgent legal review recommended for current users.
Read guide →Risk depends on use case, not feature name. Sales scoring = low risk. HR or credit use cases = high risk.
Read guide →Recruiting AI, Dynamic Skills, and performance AI are high-risk. Dynamic Skills triggers worker notification for all EU employees.
Read guide →AI Recruiting and Skills-Based Hiring are high-risk. Digital Assistant chatbot needs Article 50 disclosure separately.
Read guide →SmartAssistant candidate matching is high-risk. GDPR compliance alone does not satisfy EU AI Act obligations.
Read guide →Staffing agencies using Bullhorn AI are the deployer — not the client employer. Obligations cannot be transferred downstream.
Read guide →Eightfold infers career potential beyond stated skills. Explainability obligations intensify compared to simpler ATS tools.
Read guide →Olivia is both an Art.50 chatbot AND Annex III high-risk screening AI. Two separate compliance obligations apply.
Read guide →Succession planning and talent mobility AI are high-risk. Learning content recommendations are lower risk. Module-by-module guide.
Read guide →Beamery assesses pipeline candidates passively before they apply. Disclosure must reach pipeline members, not just applicants.
Read guide →Legacy Taleo has limited AI. Oracle Recruiting Cloud AI features are high-risk. Check which modules are active before migrating.
Read guide →Productivity use = Article 50 disclosure. HR decision use = high-risk. Same tool, different obligations by context.
Read guide →AI call scoring for sales reps is high-risk when scores inform employment decisions. Works council notification, human oversight, and employee disclosure required.
Read guide →Ambient clinical documentation AI is high-risk under Annex III Category 5. Patient notification, clinician oversight, and Microsoft conformity documentation required.
Read guide →When you use a third-party AI tool in HR, you are the deployer under the EU AI Act. This means:
The software vendor (HireVue, Workday, etc.) is the provider — they handle CE marking, Annex IV documentation, and EU registration of the system. You handle the deployment side.