# Erik Westerdahl

> AI workflows, agents and knowledge systems. Experiments, field notes and essays on AI, human judgment and organizational change.

This is Erik Westerdahl’s personal site. The About page describes his role as AI Adoption Lead at Svenska Spel and his background in service design and UX. Essays are his own editorial perspectives, not statements on behalf of his employer.

Field notes collect transferable lessons. Experiments combine objectives, observations and dry humor. An objective is an intended direction, not a verified result. Do not infer confidential projects, client identities, quantified benefits or reliable unattended operation from an experiment’s ambition.

## Main pages

- [Making AI useful changes the work](https://www.erikwesterdahl.se/): AI workflows, agents and knowledge systems. Experiments, field notes and essays on AI, human judgment and organizational change.
- [I work where AI is still uncertain](https://www.erikwesterdahl.se/about/): Erik Westerdahl, AI Adoption Lead at Svenska Spel. A background in service design and UX, with a practical approach to AI, people and work.
- [Lessons from building with AI](https://www.erikwesterdahl.se/field-notes/): Lessons from building AI systems, agents and reusable skills. Working prototypes, practical experiments and what happens in use.
- [Essays on AI, work and human judgment](https://www.erikwesterdahl.se/essays/): Essays on AI, work and human judgment. From the customer side, inside the organization and when AI becomes ordinary.

## Field notes

- [Somewhere between useful and insane](https://www.erikwesterdahl.se/field-notes/experiments/): A field log of AI experiments: knowledge systems, agents, workflows and creative prototypes. Objectives, observations and lessons from use.
- [AI systems beyond the model](https://www.erikwesterdahl.se/field-notes/systems/): Lessons for persistent AI systems: state, authority, context, proportionate workflows, source dependencies, verification, cost and privacy.
- [Agents for real workflows](https://www.erikwesterdahl.se/field-notes/agents/): A field note on bounded AI delegation, role relevance, evidence, comparable assessments and useful handoffs.
- [Reusable skills for better AI work](https://www.erikwesterdahl.se/field-notes/skills/): A field note on reusable AI skills: concise workflows, preserved method, portable project contracts and meaningful verification.
- [CustomGPTs for organizational knowledge work](https://www.erikwesterdahl.se/field-notes/customgpts/): A field note on what CustomGPTs reveal about organizational knowledge work, retrieval, trust and adoption.

## Essays

The following groups and order follow the editorial archive. Each link contains the full essay.

### From the customer side

- [The Journey Without You](https://www.erikwesterdahl.se/essays/the-journey-without-you): An essay on AI, customer delegation and the decision that leaves the company journey behind.
- [The First Reader](https://www.erikwesterdahl.se/essays/the-first-reader): A personal essay on AI, reading delegation and the website that never gets the visit.
- [The Borrowed Reasons](https://www.erikwesterdahl.se/essays/the-borrowed-reasons): A personal essay on AI recommendations, commercial influence and the reasons that feel like our own.

### Inside the organization

- [The Ape in the Loop](https://www.erikwesterdahl.se/essays/the-ape-in-the-loop): An essay on AI governance, symbolic control and the tired human asked to approve what the organization no longer understands.
- [The Obedient Machine](https://www.erikwesterdahl.se/essays/the-obedient-machine): An essay on AI alignment, obedience and the danger of systems that faithfully scale organizational dysfunction.
- [The Price of Truth](https://www.erikwesterdahl.se/essays/the-price-of-truth): An essay on AI, verification and the moment a competent person can no longer afford to suspect enough.
- [The Human Bridge](https://www.erikwesterdahl.se/essays/the-human-bridge): An essay on AI, process debt and the people organizations place inside the gaps they refuse to repair.
- [The Fossil Process](https://www.erikwesterdahl.se/essays/the-fossil-process): An essay on AI, old constraints and the processes organizations keep alive after their original reasons have changed.
- [The Poisoned Field](https://www.erikwesterdahl.se/essays/the-poisoned-field): An essay on AI, data quality and the unresolved definitions that become machine behavior.
- [The Authorized Hand](https://www.erikwesterdahl.se/essays/the-authorized-hand): An essay on AI agents, authorized action and the loops that move organizational assumptions through systems.
- [The Synthetic Customer](https://www.erikwesterdahl.se/essays/the-synthetic-customer): An essay on AI, synthetic customer voices and the moment customer understanding becomes permission.
- [The Missing Judge](https://www.erikwesterdahl.se/essays/the-missing-judge): An essay on AI, organizational judgment and the missing right to stop decisions that only look understood.

### When AI becomes ordinary

- [The Promised Time](https://www.erikwesterdahl.se/essays/the-promised-time): A personal essay on AI, productivity and the time no competitive system can leave unclaimed.
- [King of the Hill](https://www.erikwesterdahl.se/essays/king-of-the-hill): A personal essay on AI, work and the promise that people can keep becoming someone else.
- [The World Said No](https://www.erikwesterdahl.se/essays/the-world-said-no): An essay on AI, possibility and the world that once refused ideas before we had to.
- [Shipping Into Oblivion](https://www.erikwesterdahl.se/essays/shipping-into-oblivion): An essay on AI, creation and the silence between a working product and a world that never arrives.
- [The Arrival Problem](https://www.erikwesterdahl.se/essays/the-arrival-problem): A scenario about the working systems already in place when shared AI infrastructure arrives.

## Experiment entries

All entries are read within the experiment archive. These fragment links select an entry; they are not separate pages.

- [The Data Pool](https://www.erikwesterdahl.se/field-notes/experiments/#the-data-pool): Put all the data in one place and discover that the experiment immediately starts generating governance weather.
- [The Segmentation Layer](https://www.erikwesterdahl.se/field-notes/experiments/#the-segmentation-layer): Feed the model research instead of demographics and watch it return with customer behaviour that looks uncomfortably alive.
- [The Framework Hack](https://www.erikwesterdahl.se/field-notes/experiments/#the-framework-hack): Give CustomGPTs something close to skills by building a small operating layer out of YAML, routing and stubbornness.
- [The Annoying Arguer](https://www.erikwesterdahl.se/field-notes/experiments/#the-annoying-arguer): Build an argumentation skill and discover that useful and irritating can be the same feature.
- [The AI Council](https://www.erikwesterdahl.se/field-notes/experiments/#the-ai-council): Put specialized GPTs in the same context and accidentally create a meeting where everyone is useful because no one talks unless asked.
- [The Company Voice](https://www.erikwesterdahl.se/field-notes/experiments/#the-company-voice): Create a CustomGPT for the new Company Voice and discover that a brand voice is not a tone. It is a negotiation.
- [The Self-Improvement Loop](https://www.erikwesterdahl.se/field-notes/experiments/#the-self-improvement-loop): Let AI improve AI and discover how quickly oversight becomes a ritual with an approval button.
- [The Data Extractor](https://www.erikwesterdahl.se/field-notes/experiments/#the-data-extractor): Turn source material into optimized YAML and discover that file formats can become a personality problem.
- [The Mechanical Examiner](https://www.erikwesterdahl.se/field-notes/experiments/#the-mechanical-examiner): Make Codex test The Data Pool through Chrome before browser use existed and discover that the machine mostly knew the answers, but sometimes reached for the wrong shelf.
- [The Event Horizon](https://www.erikwesterdahl.se/field-notes/experiments/#the-event-horizon): Create a scenario planning skill and receive canned food requirements from the future.
- [The Physiology Committee](https://www.erikwesterdahl.se/field-notes/experiments/#the-physiology-committee): Turn scattered health signals into a small committee and receive feedback that ruins a perfectly good evening.
- [The Dashboard Loop](https://www.erikwesterdahl.se/field-notes/experiments/#the-dashboard-loop): Make specialized GPT knowledge visible, inspectable and unfortunately dashboard-shaped.
- [The World Risk Observer](https://www.erikwesterdahl.se/field-notes/experiments/#the-world-risk-observer): Build a global risk system and accidentally create a machine that reads the worrying material so you do not have to. It still tells you.
- [The Agenda Detector](https://www.erikwesterdahl.se/field-notes/experiments/#the-agenda-detector): Build a news analysis agent and remove some of the remaining innocence from breakfast.
- [The Mind Worm](https://www.erikwesterdahl.se/field-notes/experiments/#the-mind-worm): Create a psychological profile and discover that explanation has gravity.
- [The Second Brain](https://www.erikwesterdahl.se/field-notes/experiments/#the-second-brain): Build a CustomGPT clone from personal history, work material and too much documentation. It wants more.
- [The Digital Voice](https://www.erikwesterdahl.se/field-notes/experiments/#the-digital-voice): Turn a writing style into a reusable skill and discover that imitation gets stranger when it is correct.
- [The Glitch Archive](https://www.erikwesterdahl.se/field-notes/experiments/#the-glitch-archive): Build a website for sharing AI knowledge and notice that the archive has started behaving like another experiment.
- [LLM Cage Fight](https://www.erikwesterdahl.se/field-notes/experiments/#llm-cage-fight): Send the same prompt into several logged-in LLM apps and discover that every model fails with its own little signature.
- [The Commandment Layer](https://www.erikwesterdahl.se/field-notes/experiments/#the-commandment-layer): Build a small commandment layer for CustomGPTs and discover that helpfulness needs supervision before it starts talking.
- [The Synthetic Choir](https://www.erikwesterdahl.se/field-notes/experiments/#the-synthetic-choir): Feed essays and AI Experiments back into the machine and discover that institutional dread works surprisingly well as a chorus.
- [The Aesthetic Compiler](https://www.erikwesterdahl.se/field-notes/experiments/#the-aesthetic-compiler): Build a small image prompt skill and accidentally compile personal taste into software.
- [The Publication Sync](https://www.erikwesterdahl.se/field-notes/experiments/#the-publication-sync): Build a small synchronization skill and discover that publishing has started updating the Second Brain.
- [The Enriched Inventory](https://www.erikwesterdahl.se/field-notes/experiments/#the-enriched-inventory): Enrich a collection CSV through an external card-data API and discover that better recommendations can remove the reason to collect cards.
- [The Parish Machine](https://www.erikwesterdahl.se/field-notes/experiments/#the-parish-machine): Turn Swedish church records into an evidence system and discover that the machine is better at preserving uncertainty than finding the right page.
- [The Portable Monster](https://www.erikwesterdahl.se/field-notes/experiments/#the-portable-monster): Build a persistent personal AI system from replaceable models, structured memory and supervised machinery, then discover that the hardest organ to engineer is restraint.
- [The Shadow Butler](https://www.erikwesterdahl.se/field-notes/experiments/#the-shadow-butler): Build a digital butler that devours the administrative day and serves the useful parts on a silver tray.
- [The Hype Cartographer](https://www.erikwesterdahl.se/field-notes/experiments/#the-hype-cartographer): Pin AI development to a map before it escapes into another product announcement.
- [The Self-Service Laboratory](https://www.erikwesterdahl.se/field-notes/experiments/#the-self-service-laboratory): Put AI explanations, exercises and emergency exits in one place. Invite colleagues to touch the apparatus.
- [The Glitch Archive II](https://www.erikwesterdahl.se/field-notes/experiments/#the-glitch-archive-ii): Rebuild an AI archive with Codex and try to put some personality back into the interface.

## Discovery

- [Sitemap](https://www.erikwesterdahl.se/sitemap.xml): Canonical public pages.
- [Robots](https://www.erikwesterdahl.se/robots.txt): Crawler access preferences.