Development project · a trial for an invited group

Segullinn

Understand what your people are saying. Find the next step.

Segullinn (Icelandic for “the magnet”) is a short survey, 16 statements and two open questions, that shows where your people experience clarity, fairness and trust most and least, and which questions are worth discussing next.

The trial survey is currently in Icelandic only.

The button opens an email to Kjartan with four short questions. You will get a reply within 2 working days. Nothing opens? Write to kjartan@verslanagreining.is.

The trial is free and for a small, invited group. Participants are chosen from those who express interest. In the trial each respondent needs a claude.ai account.

Have you been invited? Open the trial (in Icelandic; invitation and sign-in required)

Where things stand

What is available, and what is not

What Segullinn is now

A development trial testing a short survey meant to help managers get indications of how employees experience their daily work.

What a manager gets in the trial

A short survey and a summary of the answers. Once there are enough answers, results by factor and the main themes from open answers appear. The summary can point to questions worth looking at next. Respondents need a claude.ai account.

What the results can tell you

They are indications for discussion, not a confirmed analysis of causes. Segullinn does not rule on misconduct, bullying, health problems or the responsibility of individual employees. The questions have not yet been scientifically tested.

Vision only

Ideas under consideration, not services available now:

  • A full system review and interviews
  • Comparison between workplaces
  • A workplace index
  • A recognition
  • “Ask the team”

Read about the vision

The trial is free. Later pricing has not been decided. Taking part in the trial is by invitation. You can register interest or get in touch to discuss whether it suits your workplace. The email address is kjartan@verslanagreining.is.

How the trial works

Four steps in the trial

  1. A manager sets up a survey

    16 short statements on clarity, fairness, trust, leadership, collaboration and other factors, plus two open questions.

  2. Respondents get access and a link

    Each respondent needs a claude.ai account. Before the survey is sent, respondents get a link to claude.ai and are asked to create an account if they do not have one, in a browser or in the Claude app. The survey is then shared by link. It asks for neither names nor email addresses. The survey is currently in Icelandic only.

  3. Responses are followed

    The manager sees how many have answered, not what any individual answered.

  4. Summary

    Once there are enough answers: results by factor, the main themes from open answers and questions worth looking at next.

Suits workplaces from about 10 employees. Answering takes about five minutes. The measurement can be repeated later with a new survey. Re-measurement is not automatic.

Example summary

What the summary could look like

The image shows the kind of summary the trial produces. The data is made up.

  • Results by factor. Where the group scores strongest and weakest.
  • Main themes from open answers. Summarised by AI. The answers themselves are not shown.
  • Questions to look at next. For discussion in the team, not instructions.
Summary · ExampleConcept image · made-up data

Results by factor

Collaborationstrongest
Fairnessweakest
Claritywithin the margin

Main themes from open answers

Uncertainty about who decides when shifts change. Information about changes arrives late.

Questions to look at next

1
Do people know who may decide what?For discussion, not a finding
2
What is actually rewarded?For discussion, not a finding
3
How does information about changes reach people?For discussion, not a finding
Concept image with made-up data. No real companies, answers or figures.

Confidentiality and data

What the manager sees, and what they do not

In the trial as it is now

The manager sees

  • How many have answered.
  • Results by factor once at least five people have answered.
  • Main themes from open answers once at least seven have answered, not the answers themselves.

The manager does not see

  • Individual employees’ answers.
  • Results for groups that are too small. These are marked “not showable”.
  • No names in the survey. The survey asks for neither name nor email address. Respondents do, however, sign in to claude.ai to reach the trial. The trial records whether an account has answered, so nobody answers twice, but that record is not linked to the answer.
  • How the summary is made. Numerical results are calculated with fixed formulas, using the same model as the Segullinn calculation model. Themes from open answers are summarised by AI (Claude, by Anthropic). It is asked not to quote verbatim or mention anything that could identify people, but this cannot be fully guaranteed.
  • Where the data is. Trial data is stored on claude.ai (Anthropic). The person running the trial can technically see the answers but does not look at them except to fix errors.
  • Open text. What is written in open answers can give clues about who answered, especially in small groups.
  • Not full anonymity. Technical anonymity has not been verified, so it is not promised.
  • Not sold or passed on. The data is neither sold nor passed on to others.

Responsibility and privacy

Segullinn is operated by Straumrót ehf. The trial processes survey answers and whether an account has answered. The data is stored on claude.ai (Anthropic) and used only to produce a summary for the workplace concerned. Answers are deleted no later than 6 months after the trial ends. You can ask for data to be deleted sooner by writing to kjartan@verslanagreining.is.

Principles for the full platform

Why Segullinn

The number tells you what. It does not always tell you why.

Employee surveys show trends, and many do so well. Some also compare workplaces and look for explanations. In the longer term, Segullinn aims to connect employee experience, possible explanations, next steps and follow-up in one process.

Side route · idea under consideration

Just need to ask the team about one project?

The idea is that you could create a short survey about a project or way of working and get simple counts, with no scores or analysis.

The “Ask the team” tool does not exist yet. It would be kept separate from Segullinn’s survey, with no analysis or index, and the answers would not be used in any comparison.

Register interest in “Ask the team”

How it would work
  • Four steps. A title and short introduction, questions (about ten at most), a preview and one link to share, and finally a count of answers.
  • One scale. Strongly disagree, Disagree, Neutral, Agree, Strongly agree, Not applicable. Open questions can be added.
  • Number of answers, not response rate. A link can be forwarded, and the tool does not confirm that each person answered exactly once.
  • No respondent login. No name or email address is asked for. The manager logs in to create a survey and see results.
  • A minimum of five answers per question. If a question has too few answers when the survey closes, the tool says “not shown, too few answers”, not that nothing was found.
  • Open answers can identify. Respondents are told this before they answer, and open questions can be skipped. A five-answer minimum does not make a written answer untraceable.
  • About work, not people. A reminder not to ask about named individuals, health, bullying, harassment or disciplinary action. An automatic warning helps but is no guarantee.
  • Retention. The proposal is to delete answers six months after closing. This is not a confirmed rule.
Kjartan Örn Sigurðsson

About the author

Who is behind Segullinn?

Kjartan Örn Sigurðsson has held senior leadership roles for more than 25 years in Iceland, the rest of the Nordics, Europe and the USA. He was CEO of Ormsson from 2021 to 2026, founded and built SRX, and led the restructuring of Egilsson after the 2008 financial crisis. Before that he ran sales and marketing for STRAX in Europe from London.

He has served on the boards of listed and private companies, including the Confederation of Icelandic Enterprise (SA) and the Federation of Trade and Services (SVÞ), and was a municipal councillor in Álftanes and Garðabær. He holds an MBA from the University of Iceland.

Segullinn grew out of that experience: problems at work more often lie in the rules and systems than in the people.

Let’s talk

Want to discuss the trial, review the method or talk about working together?

Register interest in taking part in the trial, give feedback on the method or discuss possible collaboration. Get in touch with Kjartan Örn Sigurðsson.