{"wiki":{"id":1042,"slug":"es-s1","entity_type":"droid","entity_id":"es-s1","title":"Estimation Accuracy","prose_md":"## About\n\nEstimation Accuracy helps [[agt_021]] improve future budgets. After an initiative finishes, it compares the original budget with actual costs, calculates the variance and average cost per completed step, and checks whether the underlying estimation formulas consistently overestimate or underestimate costs.\n\nIt recommends adjustments when actual costs differ from estimates by more than 30% and logs its findings, but it cannot change budgets, coefficients, or historical records itself. This is a uniquely named droid within the reusable intelligence family, which has 25 droids run across several agents. It is pure software and makes no Claude calls.","namesake_json":"{\"engine\": \"software\", \"model\": null, \"claude_purpose\": null, \"family\": \"intelligence\", \"family_size\": 25, \"is_unique\": true}","profile_pic_path":"avatars/droid/es-s1.png","source_hash":"3cc605f28ae8df981ab3ed0bb29c85a242b1a66c9cba2f6629f839cd4d2f9f00","status":"active","last_generated_at":"2026-08-04 01:39:35","created_at":"2026-07-27 03:30:50","updated_at":"2026-08-08 03:26:28"},"facts":{"id":"es-s1","name":"Estimation Accuracy","machine_name":"ES-S1","owner":"agt_021","function":"Post-initiative budget vs actual cost analysis","process_type":"triggered","schedule":"On demand","process_name":"estimation_accuracy","avatar_color":"#eab308","created_for":"Midas must ensure the organization stays within its means by managing budgets and improving cost estimates. A critical part of this work is learning from completed initiatives: measuring how accurate past budgets were and using that data to refine future estimates. ES-S1 covers the post-initiative analysis — comparing actual costs to original budgets to reveal estimation gaps.","purpose":"When an initiative completes, ES-S1 gathers the actual costs incurred and measures them against the original budget to show Midas where estimates were accurate and where they missed. By calculating the average cost per step, it reveals whether the underlying estimation coefficients (the formulas that predict costs) are calibrated correctly. This feedback loop lets Midas adjust those coefficients so future budgets get more accurate.","duties":"- Retrieves completed initiative budgets, planned steps, and actual recorded costs from the database\n- Calculates variance (budget minus actual) and average cost per step completed\n- Compares actual per-step costs to the base estimation coefficient to detect systematic overage or underestimation\n- Recommends coefficient adjustments when actual costs deviate significantly (more than 30%) from the estimate\n- Logs all analysis results with timestamp for Midas's financial records","constraints":"ES-S1 only analyzes completed initiatives; it cannot forecast future costs. It recommends coefficient adjustments but does not apply them automatically. It cannot modify initiative budgets or change historical cost records.","status":"running","gender":"female","archetype":"intelligence","building_block":"cost_capture","service_override":null,"enabled":1,"task_key":"cost-center.cost_estimated","model_free":0},"page":{"type":"droid","page_class":"main","class_label":"Main page","kind_label":"Droid","kind_gloss":"","listed":true}}