{"slug": "bug-report-analyst-capacity-leaks-into-retrieval", "title": "Bug Report: Analyst Capacity Leaks Into Retrieval", "summary": "The team at TechnoXander published a bug report framing manual evidence retrieval in anti-money-laundering investigations as a high-severity workflow defect, arguing that analysts lose significant capacity assembling customer history, linked accounts and timelines by hand before they can begin analysis. The proposed fix is to retrieve evidence before an analyst opens a case, surface sources, and keep the human in charge of the decision, drawing on RAG-based AI for explainable AML investigations. The team recommends logging a week of analyst work to measure the cost before fixing it, and stresses the issue is a workflow bug rather than a model-accuracy problem.", "body_md": "Engineers know how to write a good bug report. Compliance rarely gets one. So here is one, for the most expensive defect in many AML operations.\n\nTitle: Analyst capacity drops sharply when retrieval is manual\n\nSeverity: high, but invisible on every dashboard\n\nEnvironment: an investigation team with several systems and no single case view\n\nSteps to reproduce\n\nOpen an alert.\n\nTry to find the customer’s earlier cases.\n\nTry to find linked accounts.\n\nTry to assemble a timeline.\n\nObserve how long it takes before you can start thinking.\n\nExpected: the case opens with history, links and a draft chronology.\n\nActual: the analyst assembles them by hand, then decides.\n\nWorkaround: ask a senior colleague. This works until that colleague leaves.\n\nSuggested fix: retrieve before the analyst opens the case, show sources, keep the person in charge of the decision. See the idea behind [RAG-powered AI that generates evidence](https://technoxander.com/rag-ai-aml-investigation-explainable/).\n\nNotes for triage\n\nThis is not a model-accuracy bug. It’s a workflow bug.\n\nMeasure it before fixing. A one-week log is enough: [https://www.pearltrees.com/technoxander47/item838554012](https://www.pearltrees.com/technoxander47/item838554012) shows how to sort which tasks to fix first.\n\nReproduced in a hypothetical day: [https://verificationofpayee.blogspot.com/2026/11/day-in-two-aml-workflows.html](https://verificationofpayee.blogspot.com/2026/11/day-in-two-aml-workflows.html).\n\nContext: [AML investigation capacity](https://technoxander.com/aml-ai-investigation-throughput-scaling/) and [organisational memory in AML](https://technoxander.com/organisational-memory-aml-ai-investigations/).\n\nOver to you\n\nHow do you label tickets that cost time but never break anything?\n\nWould you ship a fix that moves work from analysts to a retrieval layer, even if accuracy needs monitoring?\n\nHot take: the worst bugs are the ones with no error message. Agree, or fight me?\n\nFrom the team at [TechnoXander](https://technoxander.com/).", "url": "https://wpnews.pro/news/bug-report-analyst-capacity-leaks-into-retrieval", "canonical_source": "https://dev.to/technoxander/bug-report-analyst-capacity-leaks-into-retrieval-1007", "published_at": "2026-11-11 14:14:22+00:00", "updated_at": "2026-11-11 14:53:05.529404+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-agents", "ai-products"], "entities": ["TechnoXander"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/bug-report-analyst-capacity-leaks-into-retrieval", "markdown": "https://wpnews.pro/news/bug-report-analyst-capacity-leaks-into-retrieval.md", "text": "https://wpnews.pro/news/bug-report-analyst-capacity-leaks-into-retrieval.txt", "jsonld": "https://wpnews.pro/news/bug-report-analyst-capacity-leaks-into-retrieval.jsonld"}}