[{"data":1,"prerenderedAt":353},["ShallowReactive",2],{"post-\u002Fprotecting-organizational-memory-during-ai-transformation":3},{"page":4,"translations":200,"nav":208,"related":336,"random":345},{"id":5,"title":6,"body":7,"categories":170,"category":173,"changeHistory":173,"date":174,"description":175,"disclosures":176,"draft":179,"extension":180,"firstLiveAt":173,"image":181,"imageAlt":182,"kind":183,"lang":184,"meta":185,"navigation":186,"omitGermanLocalizationDisclosure":179,"path":187,"publishedAt":173,"readingTime":188,"rights":173,"seo":189,"seoTitle":6,"slug":190,"sources":173,"stem":190,"tags":191,"translationKey":197,"type":198,"updated":173,"__hash__":199},"posts\u002Fprotecting-organizational-memory-during-ai-transformation.md","Protect Organizational Memory During an AI Transformation",{"type":8,"value":9,"toc":161},"minimark",[10,14,17,30,35,38,47,50,53,56,60,63,66,69,72,75,78,85,88,92,95,98,101,104,107,111,114,117,125,128,131,135,138,141,147,150,153,158],[11,12,13],"p",{},"A company might move some software developers from product teams to teams that develop AI models. In that work, a developer asks a model to complete a software task, tests whether the resulting code works, and identifies where it falls short. That is what I mean here by model evaluation. The example is not limited to software. Moving a salesperson to customer support because chatbots have arrived, or moving part of a finance team into sales because automation has increased capacity, can be changes of the same order.",[11,15,16],{},"Developing AI models also requires software knowledge and judgment. But building a product and contributing to model development are different experiences in an employee’s daily work and career direction. When a company moves people into a new initiative, it needs to account for that difference, for their expectations, and for the responsibilities they leave behind.",[11,18,19,20,29],{},"In ",[21,22,28],"a",{"href":23,"rel":24,"target":27},"https:\u002F\u002Fblog.pragmaticengineer.com\u002Fthe-pulse-metas-self-inflicted-resignation-wave\u002F",[25,26],"nofollow","noopener","_blank","his account of departures after Meta’s restructuring",", Gergely Orosz reports that some experienced engineers pursued other jobs after being moved into AI work they did not want. Some employees he spoke with left despite additional equity offers intended to retain them. This account draws on employee interviews and examples the author says he verified. It is not a comprehensive measure of company-wide departures or of why every employee left.",[31,32,34],"h2",{"id":33},"a-change-of-assignment-also-changes-a-professional-future","A change of assignment also changes a professional future",[11,36,37],{},"Someone who has spent years developing a product may find model evaluation compelling. Someone else may feel removed from product decisions. Moving both people into the same role does not create the same professional opportunity for each of them.",[11,39,40,41,46],{},"Pascal Bornet describes ",[21,42,45],{"href":43,"rel":44,"target":27},"https:\u002F\u002Fpascalbornet.substack.com\u002Fp\u002Fefficient-and-empty-a-conversation",[25,26],"a similar tension in his account of a conversation with an experienced bank relationship manager whose role narrowed after automation",". She was not struggling to use the new technology; she was struggling to understand which part of the work still belonged to her. The example shows why a productivity gain is not the same as preserving an employee’s sense of contribution and professional direction.",[11,48,49],{},"Employees want to know how they will use what they already know in the new role, what they will learn, and how their work will be assessed. They also need to understand where the role fits in their professional future. Keeping their salary unchanged does not resolve every concern when those questions remain open.",[11,51,52],{},"When company priorities change, employees may be asked to take on new responsibilities. But the company needs to explain why it chose a particular person and how the transition will work. If it does not yet know how long the work will last, that uncertainty should be shared too.",[11,54,55],{},"It is reasonable to ask people to learn new skills. It is different to expect them to find the time and support for that learning on their own. That passes the cost of the transition to the employee.",[31,57,59],{"id":58},"what-a-company-knows-is-broader-than-its-documentation","What a company knows is broader than its documentation",[11,61,62],{},"Return to the developer who moves into model evaluation. The company’s headcount stays the same because no one has left. But if that person can no longer spend time with the former team and nobody takes over their responsibilities, the team may struggle with work that depended on that person’s judgment.",[11,64,65],{},"You can learn how a system works from its code and documentation. It can be harder to learn why it was built that way.",[11,67,68],{},"In a software team, an experienced person may know which customer problem an apparently unnecessary check prevented in the past. Someone else may see why a technically clean change will create extra work for operations. That knowledge is not always kept in a separate document. It is embedded in daily decisions.",[11,70,71],{},"When assessing a senior developer’s contribution, a company needs to consider that judgment alongside the code they write. Stopping a team before it starts solving the wrong problem, noticing that a change will affect another system, or teaching a junior colleague which question to ask are parts of the work too.",[11,73,74],{},"The same kind of knowledge is not limited to software teams. An experienced customer-service representative knows which objection cannot be resolved with the standard answer. A production worker may recognize that an ordinary-looking deviation signals a larger problem.",[11,76,77],{},"As these people change roles, the company also needs to preserve their ability to use and transfer what they know. If handover consists only of passing along files, the reasons behind decisions can disappear. An AI system with access to the documents cannot restore reasons that were never recorded.",[11,79,80],{},[81,82],"img",{"alt":83,"src":84},"An experienced employee shares the decision context behind the files with the person taking over the work","\u002Fimages\u002Finline-organizational-memory\u002Ftacit-knowledge-handover.avif",[11,86,87],{},"This loss does not require the employee to leave the company. If nobody consults them anymore, or their new role leaves no time for it, the company can become unable to use what they know.",[31,89,91],{"id":90},"uncertainty-changes-how-the-remaining-team-works","Uncertainty changes how the remaining team works",[11,93,94],{},"The effects of restructuring are not limited to people who change roles or leave. The employees who remain also draw conclusions about their own future from what they see.",[11,96,97],{},"If someone believes the reason for their work can be devalued without explanation, they may be more cautious about taking long-term responsibility. When people do not know where the next change will leave them, looking at other jobs becomes a reasonable option. The company can then risk losing people it hoped to keep.",[11,99,100],{},"Employee concern should not be read only through productivity. Job security, income, and professional future affect people’s lives directly. Whether management takes those concerns seriously also shapes trust in its commitments.",[11,102,103],{},"When trust declines, the way people work day to day can change. Rather than raising problems early, making time for colleagues, or taking ownership of an improvement with an uncertain outcome, people may focus on protecting themselves.",[11,105,106],{},"Retaining someone is not the same as continuing to benefit from what they know. Extra compensation may help with the first. The second also needs meaningful responsibility, fair assessment, and a voice that is taken seriously.",[31,108,110],{"id":109},"lower-payroll-does-not-describe-the-full-cost","Lower payroll does not describe the full cost",[11,112,113],{},"AI may reduce the labor required for some tasks. That can affect staffing needs. A sound calculation also needs to include how the work that remains will be carried out.",[11,115,116],{},"When experienced people are moved out of a team, more checking, consultation, or relearning may be needed. A problem once resolved in a short conversation can turn into meetings involving several teams. If errors are noticed later, they may cost more to correct.",[11,118,119,120,124],{},"These are not losses that must follow every reduction in staffing. They are assumptions that need testing within the expected savings. I explored why AI’s value should not be measured only by producing more work in ",[21,121,123],{"href":122},"\u002Fhow-can-companies-create-value-from-ai","How Can Companies Create Value from AI?",".",[11,126,127],{},"Measures should therefore include completion time, rework, and service quality alongside payroll cost. A company should also assess how many people understand critical issues and whether work can continue without them.",[11,129,130],{},"If today’s costs fall while the company’s ability to solve problems falls too, it may not yet have counted the future cost.",[31,132,134],{"id":133},"design-the-transition-with-the-people-who-know-the-work","Design the transition with the people who know the work",[11,136,137],{},"A responsible transformation makes clear not only which jobs will change, but also which knowledge and responsibilities will be protected. It needs the knowledge of the people doing the work when those decisions are made.",[11,139,140],{},"Before roles change, a company can map current responsibilities. The reasons behind critical decisions can be recorded. Letting the new owner work alongside the experienced employee for a time can reveal missing knowledge through real work.",[11,142,143],{},[81,144],{"alt":145,"src":146},"The experienced employee and the new owner shape an AI-supported workflow together using real work examples","\u002Fimages\u002Finline-organizational-memory\u002Ftransition-designed-with-experts.avif",[11,148,149],{},"Employees moving into a new role need a concrete transition plan as well. Expectations, learning time, and support should be explicit. Assessing someone against their old performance measures while they learn a new job can put them at a disadvantage before they have begun.",[11,151,152],{},"Management may not be able to answer every uncertainty immediately. It can still say honestly which decisions are final, which remain open, and when they will be reviewed again.",[11,154,155,156,157],{},"One thing worth protecting during an AI investment is employees’ belief that they can contribute to the company’s future. In ","Where Should People Be in the Loop While AI Does the Work?",", I discussed why people need to be part of the learning loop, not merely the final checkpoint. The knowledge of experienced people can help create a better way of working. Excluding them from decisions and then trying to reconstruct that knowledge later may be an expensive way to learn.",[11,159,160],{},"When a transformation plan is approved, it is not enough to describe what the new system will do. The company also needs to know who will retain the knowledge required to use it well, notice its errors, and improve it.",{"title":162,"searchDepth":163,"depth":163,"links":164},"",2,[165,166,167,168,169],{"id":33,"depth":163,"text":34},{"id":58,"depth":163,"text":59},{"id":90,"depth":163,"text":91},{"id":109,"depth":163,"text":110},{"id":133,"depth":163,"text":134},[171,172],"ai","business",null,"2026-09-11","Why experience, organizational memory, and employee trust need protecting when an AI-led restructuring changes how a company works.",{"aiUse":177,"aiNote":178},"ai-assisted","Evren Bal developed the main idea and initial draft of this article. AI assisted with source review, structuring the text, improving the prose, and successive rounds of revision. Evren Bal reviewed and approved the final text after multiple revisions.",false,"md","\u002Fimages\u002Fhero\u002Fai-company-memory-role-transformation.avif","A sparse editorial diagram shows company knowledge carried from an existing role into a new AI-supported workflow","Analysis","en",{},true,"\u002Fprotecting-organizational-memory-during-ai-transformation",7,{"title":6,"description":175},"protecting-organizational-memory-during-ai-transformation",[192,193,194,195,196],"AI","organizational-transformation","organizational-memory","employee-experience","work-design","ai-transformation-organizational-memory","post","RHNfniA44Czyl-9CP-e2jpRsdxUgnvC51DKwXNDMM_c",{"en":201,"tr":202,"de":205},{"path":187,"title":6},{"path":203,"title":204},"\u002Ftr\u002Fyapay-zeka-donusumunde-sirketin-hafizasi-olan-calisanlari-korumak","Yapay Zekâ Dönüşümünde Şirketin Hafızası Olan Çalışanları Korumak",{"path":206,"title":207},"\u002Fde\u002Funternehmenswissen-bei-ki-transformationen-schuetzen","Unternehmenswissen bei KI-Transformationen schützen",{"prev":209,"next":212,"others":215,"lucky":333,"readingTime":188},{"path":210,"title":211},"\u002Fbringing-english-back-into-my-working-day","Bringing English Back Into My Working Day",{"path":213,"title":214},"\u002Fhow-ai-changes-the-experience-gap","Can a Junior Who Uses AI Well Outperform a Senior Expert?",[216,219,222,225,228,231,234,237,240,243,246,249,252,255,258,261,264,267,270,273,276,277,280,283,286,289,292,295,298,300,303,306,309,312,315,318,321,324,327,330],{"path":217,"title":218},"\u002Fai-made-code-cheap-verification-is-still-expensive","AI Made Code Cheap. Verification Is Still Expensive.",{"path":220,"title":221},"\u002Fpesintaksit-cash-vs-installments-a-turkish-inflation-aware-payment-comparison-tool","Cash or Installments? – The Story Behind PeşinTaksit",{"path":223,"title":224},"\u002Fredar-ai-powered-summaries-for-kap-disclosures-and-open-sources","Redar: AI-Powered Summaries for KAP Disclosures and Open Sources",{"path":226,"title":227},"\u002Fai-visibility-illusion-bing-citation-share","The AI Visibility Illusion: What Bing's Citation Share Data Actually Reveals",{"path":229,"title":230},"\u002Fllms-txt-was-never-the-point","llms.txt Was Never the Point",{"path":232,"title":233},"\u002Fan-seo-experiment-in-a-low-competition-serp-with-google-maps-and-openai","Building camiler.org: A Programmatic SEO Experiment with Google Maps and OpenAI",{"path":235,"title":236},"\u002Faccessing-know-how-is-not-the-same-as-creating-it","Accessing Know-How Is Not the Same as Creating It",{"path":238,"title":239},"\u002Fwhy-im-building-rankextension-making-google-search-console-actually-make-sense","What Happened to RankExtension?",{"path":241,"title":242},"\u002Fthe-threshold-collapsed-to-zero","The Threshold Collapsed: What ProductLog Taught Me About Building in Public",{"path":244,"title":245},"\u002Fturkeys-first-real-time-mystery-shopping-reporting","Turkey's First Real-Time Mystery Shopping Reporting Platform",{"path":247,"title":248},"\u002Fbank-account-api-integration","Integrating One Bank Is Easy. Keeping Forty Running Is Not.",{"path":250,"title":251},"\u002Fkeeping-customers-happy-isnt-enough","Keeping Customers Happy Isn’t Enough. You Have to Follow Up.",{"path":253,"title":254},"\u002Fwhy-sales-and-other-departments-keep-clashing","Why Sales and Other Departments Keep Clashing",{"path":256,"title":257},"\u002Fmanaging-technology-and-transforming-the-business-are-not-the-same","Managing Technology and Transforming the Business Are Not the Same Thing",{"path":259,"title":260},"\u002F1m-impressions-per-month-0-revenue-a-programmatic-seo-post-mortem","1M Impressions per Month, $0 Revenue: A Programmatic SEO Post-Mortem",{"path":262,"title":263},"\u002Fwriting-with-ai-means-thinking-with-your-archive","Writing with AI Means Thinking About Your Archive, Too",{"path":265,"title":266},"\u002Fis-your-ai-system-delivering-business-results","Is Your AI System Actually Delivering Business Results?",{"path":268,"title":269},"\u002Fai-turned-a-refactor-i-wouldnt-do-into-a-one-hour-job","With AI, a Refactor I Wouldn't Bother With Took One Hour",{"path":271,"title":272},"\u002Fai-assisted-rest-api-development","Preserving API Quality in AI-Assisted Development",{"path":274,"title":275},"\u002Fthe-era-of-the-previous-vibe-coder-begins","The Era of the \"Previous Vibe Coder\" Begins: The Invisibility of Clean Code and the Technical Debt Bill of AI",{"path":213,"title":214},{"path":278,"title":279},"\u002Fproductlog-the-platform-i-built-for-myself-first","ProductLog: The Platform I Built for Myself First",{"path":281,"title":282},"\u002Fbuild-in-public-2-0","Build in Public in the AI Era: What to Share and What to Keep Private",{"path":284,"title":285},"\u002Fhow-llms-identify-experts","What Makes an LLM Recommend Someone as an Expert?",{"path":287,"title":288},"\u002Fone-victory-several-defeats","One Victory, Several Defeats",{"path":290,"title":291},"\u002Fhow-i-built-a-modern-infrastructure-using-open-source-tools-and-the-power-of-cloudflare","Why I Run camiler.org on a Single VPS",{"path":293,"title":294},"\u002Fwhen-process-automation-actually-needs-ai","When Does Process Automation Actually Need AI?",{"path":296,"title":297},"\u002Fis-your-data-ready-for-ai","Is Your Data Ready for AI? Start With the Decision It Must Support",{"path":122,"title":299},"How Can Companies Create Value From AI?",{"path":301,"title":302},"\u002Ffrom-asking-questions-to-delegating-work","From Asking Questions to Delegating Work: What the Codex Study Shows",{"path":304,"title":305},"\u002Fgoogle-generative-ai-data-ai-citation-timing","AI Visibility Dropped Before Search: The Google Data That Changed My Theory",{"path":307,"title":308},"\u002Fai-value-is-not-always-more-work","AI's Value Is Not Always More Work",{"path":310,"title":311},"\u002Ftesting-a-button-treating-an-entire-website-redesign-as-a-sure-thing","Testing a Button, Treating an Entire Website Redesign as a Sure Thing",{"path":313,"title":314},"\u002Fvoice-agent-system-deployment-constraints","What to Test Before Investing in a Voice Agent",{"path":316,"title":317},"\u002Fraising-children-in-the-age-of-artificial-intelligence","Raising Children in the Age of Artificial Intelligence",{"path":319,"title":320},"\u002Fif-ai-handles-the-execution-who-sets-the-strategy","If AI Handles the Execution, Who Sets the Strategy?",{"path":322,"title":323},"\u002Fwhen-is-one-ai-model-enough","When Is One AI Model Enough?",{"path":325,"title":326},"\u002Fhow-much-authority-should-ai-have","How Much Authority Should You Give an AI System?",{"path":328,"title":329},"\u002Fdo-you-know-how-dependent-your-company-is-on-ai","Do You Know How Dependent Your Company Is on AI?",{"path":331,"title":332},"\u002Fcomprehension-debt-the-bill-comes-due-alone","Comprehension Debt: The Bill Comes Due Alone",{"path":334,"title":335},"\u002Fhealthcare-ai-chatbot-triage-medical-device","AI in Healthcare: The Difference Between Chatbots, Triage, and Medical Devices",[337,341,343],{"path":338,"title":339,"date":340},"\u002Fwhen-ai-changes-work-training-employees-is-not-enough","When AI Changes Work, Training Employees Is Not Enough","2026-09-08",{"path":307,"title":308,"date":342},"2026-09-06",{"path":301,"title":302,"date":344},"2026-09-01",[346,350,352],{"path":347,"title":348,"date":349},"\u002Fhow-to-do-content-pruning-a-real-world-case-study","How to Do Content Pruning: A Real-World Case Study","2026-08-27",{"path":268,"title":269,"date":351},"2026-09-02",{"path":322,"title":323,"date":344},1789077530584]