The research collection combines Grok-generated briefings and separately checked source digests. Original dates, publisher links, methods and limits remain visible. Repeated citations are not independent corroboration. AI + JJ applications are editorial interpretation.
- AI in municipal HR: responsibility starts in the work
- A quarter exposed: distinguish tasks from whole jobs
- Public-service staffing: scenarios create more choices
- SCB's 46 education groups: from tables to a local question
- 15 percent in customer support: learning and quality vary
- Human control needs power to change a hiring decision
- AI capability includes data, leadership and everyday learning
- The AI Act: use cases, responsibility and dated guidance
- When software manages: usefulness, understanding and responsibility
- 26 minutes is self-reported time: separate adoption from benefit
- Social dialogue on AI: skills, influence and working conditions
- AI and the working environment: follow up the relief
- Me and we with AI: better ideas in a bounded experiment
- Agentic AI in practice: early experience needs careful questions
- People stay in roles while the work changes around them
- Task and skills maps become strategic infrastructure
- AI efficiency gains lead to organisational choices
- Better performance can coexist with greater work intensity
- Time saved does not automatically become more output
- A public-service AI experiment needs purpose and judgment
- Recognise learning as part of the job
- AI work has changed. Has the job description?
- Invisible work and the right to challenge AI
- Redesign the work before adding more agents
- An announcement is not general availability
- Public HR needs capability inside the organisation
- Training is spreading. Operational readiness is uneven.
- Swedish social partners: learning belongs close to work
- Municipal HR can build capability together
- A local HR assistant starts with owned guidance
- Sjöbo: include candidate experience in the recruitment workflow
- Umeå: reported benefits still need a measurement plan
- Tjörn: read the benefit calculation and the changed working practice
- Kalmar scheduling: distinguish the baseline from pilot outcomes
- DBT: task speed, quality and organisational productivity are different outcomes
- Denmark: task changes can precede earnings and hours effects
- Swedish AI oversight: an interim mandate already exists
- Plan the work before committing to a headcount target
- Agentic HR: follow the approval boundary into execution
- A procurement route is a starting condition, not an adoption result
- Public HR needs a strategy and usable data beyond fragmented pilots
- Workforce change: separate the employer's statement from an AI explanation
- Redesign the workflow and the workforce together
- Make role owners part of AI workforce redesign
- Lidingö and IMY: AI proposes redactions, a person owns disclosure
- Talt: map the care workflow before building the AI service
- Paid work-grounded Copilot: existing permissions and a wider retrieval surface
- Check actual Copilot routing settings rather than a generic cloud promise
- Göteborg’s Copilot Chat assessment shows a real governance debate
- Eneo: shared municipal development and a local service owner
- Helsingborg participates in Sambruk’s shared AI platform
- Helsingborg’s Eneo role description makes service ownership tangible
- Svea: follow project phase, prototype and operating model separately
- Name the actual Copilot tier before comparing access and cost
- Reskilling confidence and internal moves tell different stories
- Track who gains from AI and how their work changes
- Arvika gives staff different rules for public and municipal AI tools
- An HR assistant needs reliable sources and a human fallback