Blog
If AI feels like it is moving faster than your team can plan for, you are in good company. It is one of the things we hear most from benefits and reward leaders: not sure where to start, or things shift before a decision is made.
It was also the driving theme behind our second Stockholm roundtable hosted on the 4th November. We wanted to take some of the pressure off, so we brought people together to look at what AI is changing, how it works in practice on CSRD, and how to tell when AI is the right tool and when your existing setup already does the job.
Better work, not faster work
Fredrik Wetterlundh opened the morning. His view: the next era of HR is not about doing old work faster, it is about making better work possible. AI is good at what can be seen, counted and repeated. Judgement, prioritisation, accountability and relationships stay with people. Agents, which act rather than just answer, are promising but still early.
His iceberg shows what that looks like in practice. Above the surface is the work AI already does well: the routine, repeatable tasks that fill so much of a team's week. Handing that over brings real efficiency and frees up time. But most of what matters sits below the surface, and that stays with people. The point is not to do less of it. It is to give that bottom half the time and attention to be done better.
Above the surface
AI executes and accelerates
Administration · Support · Business intelligence · Meeting notes · Travel expenses · Routine processes · Data analysis · HR services · Marketing copy · Invoice management · Customer service · Reporting · Coding and system development · Policies · Accounting
Below the surface
Humans lead, use judgement and take responsibility
Strategy · Vision · Business decisions · Culture · Values · Engagement · Relationships · Negotiations · Customer meetings · Change leadership · Conflict management · Ethical and legal boundaries
AI scales tasks. People create direction, trust and momentum.
He made the point with a simple slide: the change is coming, and your role in it is a choice. You can look away, you can collaborate, or you can be the pioneer. His encouragement to HR was to be the pioneer, with teams setting their own direction and setting the bar for how AI is used well within their own function.
In his previous role as CHRO at Scandic, his answer was practical. Training teams mattered, but so did continuous learning: time set aside in diaries every week to learn and share what people were trying. People who have played with AI are far less afraid of it.
Benefits and CSRD
CSRD and ESRS, in short
CSRD is the EU law that sets out which companies must report on sustainability. The ESRS are the standards that say what goes into that report. The social standard, ESRS S1, covers a company's own workforce, which is where benefits sit.
After a short break for lunch, we turned to the practical side, starting with a subject everyone in the room is facing: CSRD. Anna Horvath, Head of Rewards and HR Compliance at ASSA ABLOY, took us through the two paragraphs in the revised ESRS social standard that matter most to benefits professionals. S1-10 asks whether employees lack protection against loss of income across four life events: sickness, unemployment, work injury and disability, and maternity. S1-14 asks what share of employees are entitled to family-related leave. Neither asks what you spend. Both ask what you can prove.
ESRS S1-10
Do your employees have social protection?
One simple question, backed by complex evidence. Public schemes and employer benefits both count.
Sickness
Public or employer protection against loss of income.
Unemployment
Protection from the start of employment.
Work injury and disability
Income protection after injury or acquired disability.
Maternity
Income protection during maternity leave.
Where employees lack cover, disclose the country and the life event.
Answering them is complex for any multinational, and harder still across many regions. At ASSA ABLOY the picture is around 500 legal entities across more than 70 countries, with policies sitting in local documents and different languages. Gathering data on that scale, and keeping it correct and up to date, is a mammoth task. Anna advised thinking legal entity, not country, and treating a country-by-country table as the minimum. The test that matters is evidence: auditors expect it to be to hand. She also described how they plan to use AI to capture and structure this data.
ESRS S1-14
Who is entitled to family-related leave?
The measure is entitlement, not take-up. Your policies and agreements are the evidence.
Four leave types
Maternity, paternity, parental and carers' leave.
One metric
Entitled employees divided by total employees, shown as a percentage.
Evidence basis
Law, policy, agreement, contract or collective bargaining agreement.
Policy evidence sets entitlement. Employee data sets the denominator.
Where AI sits on the HR list
Start with the problem, then compare the options side by side: build in-house, use the tools your providers already offer, or bring in specialist AI. Anna's thinking gives a simple test:
- Is the work gathering, translating and structuring scattered information? Move AI up the list.
- Is the process already controlled, automated and working? Move it down. Replacing it so you can say you use AI adds risk and no value.
- Who is accountable for the answer? AI can recommend. A named person owns the outcome.
Done well, the same structured picture of benefits that answers CSRD also shows spend, coverage and renewals, while local teams keep their own decisions.
Thanks to our speakers
Thank you to Fredrik Wetterlundh and Anna Horvath for giving their time and being so open about what they are working through, and to everyone who joined us in Stockholm.
If you would like to join events like this in future, message us at hello@originbenefits.com.