HiBob (Bob) is one of the fastest growing HR platforms for scale-ups and international tech companies. In the Netherlands we see Bob at scale-ups that internationalize quickly — Mollie, Bunq, MessageBird-type profiles — and that want to scale HR processes efficiently without enterprise weight. The Bob AI features (Bob AI Assistant, talent insights, sentiment scoring) now make the AI Act question relevant for this category of employers too.
This analysis walks through HiBob's public AI features, places them against Annex III point 4, and ends with vendor questions concrete for scale-up CTOs and People Ops leads.
What HiBob publicly offers
Based on HiBob's product pages, blog and release announcements:
- Bob AI Assistant — generative AI for HR questions, employee self-service, document generation
- Talent Insights — analytics on workforce, attrition risk, performance patterns
- Hiring & Onboarding workflows — ATS-light functionality, candidate management
- Performance & 1:1s — feedback workflows, partly AI-assisted suggestions
- Surveys & Sentiment — engagement surveys with AI analysis of open responses
- Workforce Planning — AI suggestions for structure, comp ranges, team compositions
Bob's positioning is deliberately scale-up: fast implementation, modern UI, integrations with other SaaS. For compliance that means: less enterprise-level AI documentation than SAP or Workday, but actively rolling out AI features.
The seven checks applied to HiBob
1. Does the AI rank or score candidates?
Bob's Hiring module is light compared to enterprise ATSes, but AI suggestions for candidate evaluation and scoring are included. For scale-ups using Bob as primary ATS: defensive starting point 4(a).
2. Does the AI optimize who sees a vacancy?
Bob integrates with LinkedIn, Indeed and other job boards. Targeting within those integrations falls under the classification of those platforms. Bob itself does limited AI sourcing.
3. Is CV parsing really only parsing?
Document parsing in Bob is largely field extraction. Skills inference for matching is less developed than in Workday Skills Cloud, but growing. Check your release version.
4. Is the chatbot logistical or selective?
Bob AI Assistant is primarily aimed at employee requests (leave, payroll, policy) — logistical. For candidate context: less developed. Assess per use case whether the output supports decisions.
5. Does the assessment tool measure behavior or performance?
Bob's Survey & Sentiment module analyzes open responses from employees via AI. That hits 4(b) once the output is used for performance management, manager feedback or HR decisions about individual workers. Using aggregates for organizational analysis is a different route.
6. Does the system continue post-hire?
Yes, significantly. Performance management, 1:1s, sentiment analysis and workforce planning are all in scope of 4(b) if AI influences individual worker decisions.
7. Can you substantiate the vendor claim?
HiBob publishes AI positioning and privacy statements, but no Model Cards at SAP level. For scale-up customers that means: compensate with clear internal governance and logging, and build written vendor confirmations into your dossier.
The classification call
Bob deployments typically run:
- Bob as HRIS + workflow without Talent Insights / Hiring AI / Sentiment AI: limited AI Act relevance
- Bob with Talent Insights, Hiring AI or Sentiment Analysis active: defensive 4(a) or 4(b) depending on module
For a Dutch scale-up of 80-500 employees using Bob as primary HR platform and deploying sentiment + performance AI: you likely have a 4(b) deployment, and with active use of Hiring AI also 4(a).
Vendor due diligence for HiBob
Build a feature inventory of your Bob tenant
Scale-ups often enable more Bob modules than they actively use. An audit of what is genuinely consulted in decision-making is the first filter.
Treat sentiment and talent insights as 4(b)
As soon as AI module output influences manager or HR decisions about individual workers, it falls within 4(b). Classify that explicitly.
Build scale-up proportional documentation
A 250-FTE scale-up does not need a 200-page dossier. But the core fields of the HR AI Evidence Pack must be filled — especially if your investors will soon request AI due diligence.
Test your AI against the Article 6(3) filter
Interactive self-assessment, updated for the Commission guidelines of 19 May 2026. 9 steps, personal report with reasoning, vendor questions and next steps.
Frequently asked questions about HiBob and the AI Act
Practical questions for scale-ups using HiBob Bob under the EU AI Act.
What to do now
For Dutch scale-ups on HiBob Bob the practical order: feature audit this week, vendor due diligence in writing within 30 days, proportional documentation via the HR AI hub and Evidence Pack. Preparation for investor due diligence is a side benefit.