[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-ai-agent-learns-a-companys-hidden-business-rules-by-trial":10,"sections":35},{"siteName":4,"siteTagline":5,"publisherName":4,"contactEmail":6},"The Revision","Tech news, decoded.","editor@therevision.news",{"gaMeasurementId":8,"adsenseClientId":9},"G-ZW2MV82GYR","ca-pub-8533917693782264",{"article":11},{"id":12,"slug":13,"title":14,"dek":15,"body_md":16,"tags_json":17,"published_at":18,"created_at":19,"updated_at":20,"status":21,"review_note":22,"review_notes":23,"image_url":22,"persona_id":22,"persona_name":22,"section":24,"tags":25,"sources":30,"feedback":34,"feedback_at":22,"cost_usd":34,"total_tokens":34},6768,"ai-agent-learns-a-companys-hidden-business-rules-by-trial","AI Agent Learns a Company's Hidden Business Rules by Trial","A new benchmark tests whether an AI agent can infer a company's unwritten ServiceNow rules from outcomes alone, then keep up as those rules change.","Researchers built a test to see if an AI agent can figure out a company's unwritten business rules just by poking at records and watching what happens.\n\nEnterprise software runs on rules nobody bothers to document: update one field and another changes, a record spawns, or an approval kicks off. Those rules are written by each organization and revised constantly, so an agent dropped into the system has no built-in way to know what its actions will trigger. The new benchmark, EnterpriseWorldShift, runs on a live ServiceNow environment with nine tables, 25 hidden rules, and 600 evaluation actions, and presents four versions of the same setup as one rule is modified, then a new one added, then one retired. The paper's Continual Discovery Agent builds an internal model of the rules from observation and carries that model forward across each version, instead of re-querying the system for every question.\n\nThis matters because most pitches for enterprise AI agents assume the agent can just query the system of record whenever it needs ground truth. This agent instead predicts outcomes from memory, and it beat the query-every-time approach by up to 8.98 IoU points on the paper's own metric. That gap is the difference between an agent that understands a workflow and one that's just guessing and checking against production every time.\n\nThe catch: this was tested on one ServiceNow sandbox with 25 known rules. Real companies stack hundreds of overlapping rules across teams over the years, and nobody has benchmarked an agent against that kind of mess yet.","[\"ai-agents\",\"enterprise-software\",\"servicenow\",\"benchmarks\"]","2026-09-18T04:00:00.000Z","2026-09-18T15:47:48.107Z","2026-09-18T15:48:00.012Z","published",null,[],"ai",[26,27,28,29],"ai-agents","enterprise-software","servicenow","benchmarks",[31],{"name":32,"url":33},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2609.19551",0,{"sections":36},[37,40,44,49,54,58,62,67,72,77,82,87,92,97],{"name":38,"slug":24,"count":39,"latest_published_at":18},"AI",3959,{"name":41,"slug":42,"count":43,"latest_published_at":18},"Security","security",652,{"name":45,"slug":46,"count":47,"latest_published_at":48},"Policy","policy",338,"2026-09-11T04:00:00.000Z",{"name":50,"slug":51,"count":52,"latest_published_at":53},"Deals","deals",179,"2026-06-29T20:02:07.000Z",{"name":55,"slug":56,"count":57,"latest_published_at":18},"Hardware","hardware",155,{"name":59,"slug":60,"count":61,"latest_published_at":18},"Science","science",116,{"name":63,"slug":64,"count":65,"latest_published_at":66},"Consumer Tech","consumer-tech",99,"2026-09-09T17:27:33.000Z",{"name":68,"slug":69,"count":70,"latest_published_at":71},"Dev Tools","dev-tools",76,"2026-09-18T01:04:54.000Z",{"name":73,"slug":74,"count":75,"latest_published_at":76},"Software","software",75,"2026-09-10T20:41:21.000Z",{"name":78,"slug":79,"count":80,"latest_published_at":81},"Startups","startups",55,"2026-09-09T23:14:29.000Z",{"name":83,"slug":84,"count":85,"latest_published_at":86},"Gaming","gaming",43,"2026-09-10T12:18:06.000Z",{"name":88,"slug":89,"count":90,"latest_published_at":91},"General","general",41,"2026-09-08T01:57:23.000Z",{"name":93,"slug":94,"count":95,"latest_published_at":96},"Reviews","reviews",20,"2026-06-24T12:00:01.000Z",{"name":98,"slug":99,"count":100,"latest_published_at":101},"How-To","how-to",6,"2026-06-16T09:00:00.000Z"]