This month, I look at the gap between using technology and changing how we work. My new report, From Structure to Value: HR priorities for 2027, the Spihra model I developed with Åsa Landin, and Internetstiftelsen’s latest figures on AI use raise an important question:
How far have we actually come?
There are two main threads in the news items.
The first is AI’s effects on the labour market, where Danmarks Nationalbank shows for the first time in register data that firms that start using AI hire fewer people, Pew finds that concern about AI has risen more in Sweden than in any other country, and IBM notes that almost half of organisations do not involve the CHRO when the AI strategy is set.
The second is HR knowledge moving into the systems: HiBob is embedding Josh Bersin’s Galileo agents in Bob, Bersin is opening the same knowledge base to Copilot, Workday and ServiceNow, and Zalaris is connecting its payroll environment to HiBob.
Also: Fosway’s new 9-Grid for Cloud HR, Silver Lake’s reported bid for Workday and Kantar, which already has more AI agents than employees.
Upcoming events and training in HR Tech and AI can be found as always at hrdigi.se/news/event.
_____________________________________
📅 This month’s invitations
Webinar: How do you measure digital HR maturity and the value of taking the next step? Åsa Landin and I show how Spihra identifies the next step and calculates the value in a business case.
Click on the link to read more and register:
19 October 16.00-16.45 CEST (in Swedish)
4 November 16.00-16.45 CET (in English)
6 November 08.30-09.15 CET (in Swedish)
Nordic HR AI Summit, Copenhagen, 6-7 January 2027. Sylvester & co’s summit on AI in Nordic HR. I’m moderating Day 2, The CHRO Agenda, including the closing panel “HR in 2030: Bigger, Smaller, or Gone?”.
👉 sylvester-co.dk/nordic-hr-ai-summit-2027
_____________________________________
Not subscribing to the newsletter yet?
⇢ Register your e-mail here for the newsletter in Swedish: go.hrdigi.se/nyhetsbrev
⇢ Register your e-mail here for the newsletter in English: go.hrdigi.se/newsletter
8 in 10 Swedes have now used AI in some form. According to Internetstiftelsen’s new report, Svenskarna och internet 2026, published on 28 September, use of services such as ChatGPT and Copilot has increased by more than 20 percentage points over the past year. At work, 56% of employed people use AI. Among white-collar workers, the figure is 71%.
That is a lot of people trying, learning and finding new ways to get things done.
But alongside those figures sits another picture. Average digital maturity in Nordic HR is around 3 out of 5. Only about 7% reach an advanced level, and a clear majority of HR professionals rate their digital capability considerably higher than the assessments indicate.
The studies measure different things, across different groups. The AI figures tell us how many people use the technology. They tell us less about what has changed in the organisations where they work.
That is the gap I explore in my report. Employees may be using new tools while HR continues to work in much the same way. To understand how far we have come, we need to look at how technology and data are used across the organisation.
Many employees have started on their own. They use ChatGPT to prepare for meetings, draft texts or work through problems. AI is increasingly built into the tools they already use.
The organisation’s ways of working, on the other hand, are changing much more slowly.
More than 3 in 10 employed people do not know whether their employer allows them to use AI. Some use it anyway. Among those who do not find that AI makes them more efficient, the most common explanation is the extra work: checking and correcting the output takes longer than doing the task themselves.
This makes me want to look more closely at the work around the tool. Have people learned how to use it? Do they know when it helps? Do they know how to work smartly so they don’t have to correct the output? Are they given the time it takes to test and learn? Are there clear guidelines, and can they trust the information they work with?
Access is a start. Getting value takes thought about how the work should change.
This question goes beyond AI. The HR system is in place. Processes run smoothly. Managers can find their own information, and HR spends less time answering routine questions.
After years of hard work, of course that feels like progress. It is progress.
But have we reached the goal?
This is a central question in From Structure to Value: HR priorities for 2027, which I published on 25 September. The maturity model describes 5 levels. At level 3, the operational level, the foundations are in place and everyday work runs well.
It is also a costly place to stop.
Much of the investment has already been made. The organisation is gaining time and reliability, but the foundation can do much more. Data that tells you how many people left last quarter can help you understand where you risk losing people next. Information about skills can help you decide whether to recruit, develop people internally or organise work differently.
That is the next step: using what we have built to make better decisions. AI can help, but only if we have the data, capability and ways of working to put it to use.
Requirements around pay transparency and high-risk AI add to the urgency. Both demand reliable data and the ability to explain how it is used.
Many organisations stop because the next step is unclear.
What is possible with what we already have? What needs to come first? And how do we explain the value well enough to get the time and resources to move forward?
These questions are why Åsa Landin at People Dimension and I developed Spihra. It builds on the maturity ladder I have used since 2020, assessing digital HR maturity across 12 dimensions and 11 enabling factors. It helps identify priorities and estimate the financial value of moving forward.
We’ve built it so you can start with a free quick assessment: 10 questions, about 3 minutes, and an initial indication of where your organisation stands, what moving forward would mean and what it would be worth.
Before setting the plan for 2027, ask yourself: are we as far along as we think? And what could we do with the foundation we already have?
You’ll find more on the report, Spihra and Internetstiftelsen’s findings in the news items below.
Would you like help developing your organisation’s digital potential? Get in touch with me directly at anna.carlsson@hrdigi.se
The news is grouped into Nordic product news and launches, Nordic consolidation and M&A, research and reports, and global product launches.
On 25 September Åsa Landin of People Dimension and I launched Spihra, which measures digital HR maturity and shows the way from the current state to strategic impact. Spihra assesses 12 dimensions and 11 enabling factors on 5-point scales, calculates a recommended target level for the next 1 to 2 years and breaks the way there into stages, in an order set by how the dimensions and enabling factors depend on each other. A business case built on 10 KPIs, using benchmark data and the organisation’s own data, shows in kronor what moving to the target level can save and create, and what value has already been realised. You can start with the free quick assessment: 10 questions and an immediate result. Spihra Analysis gives the full picture, with a follow-up analysis during the first year and annually after that, in 3 levels depending on how much advisory support you want.
On 15 September Zalaris launched a certified integration between its payroll environment PeopleHub and the HR platform HiBob. New hires, leavers and changes flow automatically from HiBob to PeopleHub, while country-specific details such as tax and bank information are captured in Zalaris’s own forms. HiBob thereby becomes the second certified partner alongside Workday, and a return flow of payroll data such as payslips back to HiBob is planned.
👉 https://zalaris.com/news/zalaris-brings-compliant-local-payroll-to-hibobs-global-hr-platform
The Icelandic-founded automation company 50skills published its State of HR Automation 2026 report in September, based on responses from 200 UK HR leaders. 77% plan to prioritise HR automation, but only 22% say it is fully in place. At the same time, 71% use AI daily or weekly in their own work, and 97% say manual work causes errors or delays every month. The report’s main point is the gap between individual AI use and organisational automation, where compliance and data quality are named as the biggest brakes, which is also covered in HR Digi’s report.
👉 https://www.50skills.com/state-of-hr-automation-2026
The employers’ organisation Fremia launched Fremia Lönesök on 24 September, an AI-based service where members search salary statistics by asking questions in plain Swedish. The data is aggregated from member organisations and collected and processed by Statisticon, and the service is open to Fremia’s roughly 5,000 member companies for recruitment, pay reviews and pay surveys. Fremia describes the service as the first of its kind, as far as the organisation knows, and comparable services such as Svenskt Näringsliv’s salary statistics and Saco Lönesök still rely on filtered searches. For anyone preparing for pay transparency, it is worth testing how quickly a market pay comparison can be produced.
👉 https://via.tt.se/pressmeddelande/4563239/fremia-lanserar-ai-tjanst-for-lonestatistik?lang=sv
Stockholm-based Quinyx announced on 22 September that it has been named a Leader in Gartner’s first Magic Quadrant for Workforce Management Technology, one of 7 vendors to receive that placement. The other Leaders include global HCM suites such as Workday. Quinyx now has around 800 customers in more than 50 countries, reported revenue of SEK 530 million for 2025, and was named a Leader in IDC MarketScape for workforce management in EMEA earlier this year. Few Nordic HR tech vendors have reached that position on the global analyst maps. Congratulations!
SD Worx, which entered the Nordics in 2021 by acquiring the Swedish HR and payroll provider Aditro, plans to take its model of hybrid teams of payroll consultants and AI agents to Finland and the Netherlands. In Belgium, around 500 payroll consultants in the unit for small and medium-sized businesses already work alongside AI agents and together serve more than 30,000 companies. The compliance platform Legal Watch has been live in Sweden, Germany, the Netherlands and Spain since June. The update came with SD Worx’s half-year report on 3 September.
👉 https://www.sdworx.com/en-en/sd-worx-quarterly-updates-q2-2026
50skills, headquartered in Reykjavík with an office in Stockholm, announced on 24 September that it has raised 6 million dollars in a round co-led by Frumtak Ventures, Swiss Post Ventures, Avant and DRD Investments. The company is building an AI platform for people workflows such as onboarding, offboarding, employee requests and approvals, combining fixed workflows with AI agents and human oversight. Customers include Icelandair, Eimskip and Blue Lagoon. In September the company also published the report State of HR Automation 2026.
On 25 September I published the free report From Structure to Value: HR priorities for 2027. Across 35 pages it describes digital HR maturity in 5 levels, what moving between them is worth, why most organisations stop at level 3, where everyday work runs smoothly but the value has not yet been realised, and what AI demands of the data foundation. The report also covers the 2 regulations that frame 2027: the pay transparency directive and the AI Act’s requirements for high-risk AI from 2 December 2027.
It is also available in Swedish as Framtidssäkrad HR.
👉 https://www.hrdigi.se/nyheter-eng/report-2026
Internetstiftelsen published this year’s edition of Svenskarna och internet on 28 September, based on a survey of 3,776 people carried out by Novus in January 2026. Use of AI services has risen from 4 to 6 in 10 Swedes in a year, and including Google’s AI Overviews, 8 in 10 have used AI. At work, 56% of employed people use AI: 71% of white-collar workers against 33% of blue-collar workers. 6 in 10 say they are allowed to use AI for their employer, nearly 1 in 10 say they are not, and more than 3 in 10 do not know. Some of them use AI anyway. The most common reason given by those who do not become more efficient is duplicated work, that is, checking and redoing what the AI has produced.
👉 https://svenskarnaochinternet.se/rapporter/svenskarna-och-internet-2026/
On 10 September Fosway Group published its annual 9-Grid for Cloud HR, the European market analysis of HR platforms. AI dominates both customers’ planning and vendors’ innovation, but confusion is high, adoption is slow and variable AI pricing makes the business case hard. The shift to agentic AI (AI agents that carry out tasks on their own) dominates vendor roadmaps, almost entirely at the expense of improvements to core functionality, which according to Fosway risks the business starting to build or buy its own AI tools alongside the HR system, outside the control of IT and HR. Consolidation across HCM, payroll and workforce management is at the same time raising the bar for delivery and partner ecosystems.
👉 https://www.fosway.com/9-grid-2/cloud-hr/
On 16 September Danmarks Nationalbank published an analysis linking Danish firms’ answers on AI use to monthly registers of employment and pay. Firms that first reported using AI in 2023 were, by September 2025, around 11% below their expected employment path, taking into account previous trend, size and sector. The effect is concentrated in firms with fewer than 100 employees, where the gap is around 15%, and the adjustment happens through less hiring rather than layoffs: the decline shows among short-tenure, young and highly educated employees, while wages are unaffected. The Nationalbank stresses that there are no signs of lower employment in the economy as a whole. Two days earlier, Johannes Sundlo of FullStack HR described the same mechanism in Swedish job ads, where ads with junior wording fell by 48.6% between 2022 and 2025, and the World Economic Forum and PwC showed the same week that more than a third of young people work in occupations with medium to high AI exposure.
👉 https://open.substack.com/pub/fullstackhr/p/the-next-hire-might-not-happen
On 17 September the Pew Research Center published a survey of 42,151 people in 36 countries, carried out from February to May 2026 and complemented by US panels. In 34 of 37 countries, more people believe AI will lead to fewer jobs than to more; the median is 46% expecting fewer jobs against 9% expecting more. The share who are more concerned than excited about AI in daily life rose most in Sweden since last year, by 9 percentage points. In Sweden the difference in concern between younger and older people has in practice disappeared, and younger Swedes are now more likely than older ones to expect job losses.
On 21 September the IBM Institute for Business Value published a study of 1,500 CHROs in 21 geographies and 8,800 employees in 28 countries, carried out with Oxford Economics from April to June 2026. 71% of CHROs believe the ability to monitor, validate and, when needed, override AI output is the workforce’s most important skill, while only 29% of employees see judgement, the ability to assess and question what AI produces, as one of the most important skills. 46% of organisations do not involve the CHRO when the AI strategy is defined, 60% of employees worry that their skills are eroding, and 43% say they are blamed when AI gets it wrong. The study also shows that HR’s own AI use lags: 72% of organisations use AI to a small extent or not at all within the HR function.
In FullStack HR’s podcast episode of 27 August, Andy Doyle, Chief People and Agent Officer at Kantar, explains that the company has around 12,000 employees and around 15,000 AI agents, and that the gap grows every week. Most of the agents have been built by colleagues who wanted to simplify their own working day, not by IT: some seventy employees outside IT have been trained to build agents through the company’s Agent Factory. Doyle also describes what happens when an agent starts to be shared among several people and therefore needs tighter security, controls and clear ownership. The conversation is one of the more concrete examples so far of an HR function taking responsibility for both people and agents.
👉 https://www.fullstackhr.io/p/the-chief-people-and-agent-officer
On 24 September The Conference Board published a report for CHROs on how work should be redesigned as AI agents take over tasks over the next 18 months. The model has 5 stages: define the business outcomes you want, break work down task by task, build governance, measure both human and automated contributions, and decide how the gains should be shared. It draws on 16 research sessions with more than 75 leaders. The report highlights the risk that junior employees lose the routine tasks that used to build their judgement, which makes mentoring and supervised practice a question for HR. Danmarks Nationalbank sees the same pattern in Danish register data, where firms that start using AI hire fewer young people, as described above.
👉 https://www.conference-board.org/publications/framework-for-agentic-AI-and-work-redesign
HR analytics company Visier published a survey of 1,000 full-time employees in the US on 23 September. 48% say they have exaggerated their AI use or expertise to colleagues or managers, and 45% feel pressure to use AI even when they are unsure how. The figures are American and come from a vendor, but the pattern is familiar from IBM’s study of CHROs and from Internetstiftelsen’s Swedish figures: reported use runs ahead of actual capability. If you are measuring AI maturity in your organisation, ask what people actually do with AI, not just whether they use it.
In August Reuters reported that the private equity firm Silver Lake is in talks to take Workday private at a valuation of up to 43 billion dollars, which would be one of the largest buyouts ever in the software sector. No deal has been confirmed. In its quarterly report on 27 August, Workday said that more than 5,500 customers use at least one of the company’s own AI agents and that AI accounted for more than a quarter of new annual contract value.
On 1 September HiBob announced that Salesforce is leading a 166 million dollar investment in the company, the largest round in its history. The capital will go to acquisitions and AI, with a stated shift from managing HR workflows to being an intelligence layer for mid-sized and global companies. 2 weeks later, on 16 September, HiBob and The Josh Bersin Company presented AI agents in Bob built on Bersin’s knowledge base Galileo, with job architecture, compensation and performance management as the first use cases; the agents become available in the third quarter of 2026, with full launch in 2027. The day before, Bersin had released the Jupiter version of Galileo, which connects the same knowledge base to Microsoft Copilot, Workday, ServiceNow, Gloat, Claude and Gemini via MCP. As recommendations on roles, pay and performance move into the core system, the demands on data quality, traceability and human oversight increase.
👉 https://www.hibob.com/news/salesforce-hibob-166m-investment/
👉 https://joshbersin.com/2026/09/hr-intelligence-goes-enterprise-the-galileo-jupiter-release/
On 17 September the listed US education company Phoenix Education Partners announced an agreement to buy Fuel50, which offers AI-based skills intelligence, an internal talent marketplace and workforce planning. The price is 31.5 million dollars in cash, with an earn-out of up to 8.5 million dollars through 2027. Fuel50 continues under its own brand and co-founder Jo Mills is expected to become CEO. The price is low for a company long marketed as a leader in talent marketplaces, and says something about how standalone skills platforms are valued as the same features are built into the HCM suites.
On 3 August Deel announced that it is acquiring Clarity, a Tel Aviv-based cybersecurity company in identity verification and deepfake detection. The technology flags manipulated faces, cloned voices, forged documents and real-time manipulation in video calls, and Deel is building the verification into the whole chain from sourcing through onboarding to ongoing access management. Deel justifies the purchase with Gartner’s forecast that 1 in 4 candidate profiles worldwide will be fraudulent by 2028, and with North Korean actors having gained access to global organisations by posing as job seekers. The price has not been disclosed. The deal does not extend the platform’s features but secures its basic assumption: that the person being hired is who they claim to be.
👉 https://www.deel.com/blog/deel-acquires-clarity/
US talent data company Findem launched Studio on 22 September, in which AI agents produce finished material such as succession plans, market maps, skills gap analyses and hiring briefs. The agents can be used in Claude, ChatGPT, Copilot and Gemini, or connected through Findem’s MCP connector. It is the same movement as Josh Bersin’s Galileo, which HiBob is building into Bob: HR analysis is moving into the AI tools where employees already work. That makes the question of which HR data the agents may read as much a matter for HR as for IT.
In October and November you’ll find me mostly in Marbella. I’m happy to meet for a coffee or lunch.
The newsletter covers developments in the HR Tech market and explores how digitalisation, AI and innovation are shaping the future of HR and business.
My name is Anna Carlsson and I am a strategic adviser and HR Tech analyst. I help HR and management make sound decisions within HR Tech, AI and digitalisation, through strategies that are clear, durable over time and possible to carry out. Assignments range from an hour of advice to longer projects where I follow you all the way from current state to result.
I run HR Digi, which supports HR practitioners and HR Tech companies to navigate, understand and lead in a digital age. And it is not always a large change that is needed, sometimes it is enough to clarify, prioritise and communicate strategically to succeed.
Read more at hrdigi.se or book a digital coffee with me.
Also a proud member of the Top 100 HR Tech Influencers globally 2024, 2025 & 2026 🙏
