[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-83705-en":3,"doc-seo-83705-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":21,"is_downloadable":21,"audit_status":21,"page_count":22,"language":23,"language_code":24,"site_id":25,"html_lang":24,"table_of_contents":26,"faqs":27,"seo_title":13,"seo_description":14,"update_tm":28,"read_time":29},83705,4398048949847,"Eliana","https://ap-avatar.wpscdn.com/avatar/400002536579ef2da7f?_k=1778318612642679267",8,"Research & Report","Organizational Memory for Agentic Business Process Execution","LLM-based agents can automate business process execution beyond rule-based systems, yet general-purpose LLMs lack organization-specific contextual knowledge such as policies, process conventions, roles, system landscapes, and exception-handling practices. Without an enterprise-scalable way to transform fragmented BPMN, SOP, and analysis artifacts into a consistent, agent-consumable form, per-agent prompt or retrieval setups create silos. This work proposes an organizational memory, derives requirements, presents an architecture, and validates it via a procurement proof-of-concept.","arXiv :2607 .03228v 1 [ cs .AI] 3 Jul 2026  \nOrganizational Memory for Agentic Business Process Execution  \nLukas Kirchdorfer 1 , Adrian Rebmann 1 , Christian Warmuth 1 , Timotheus Kampik 1 , Theiss Heilker 1 , and Gregor Berg 1  \nSAP Signavio, Berlin, Germany  \n{[first.last}@sap.com](first.last}@sap.com)  \nAbstract. LLM-based agents offer new opportunities for automating business process execution beyond the limits of rule-based systems. However, general-purpose LLMs lack the organization-specific knowledge required for reliable execution, which is typically fragmented across humanoriented artifacts such as policies, process models, and standard operating procedures. While such knowledge can technically be encoded in individual prompts or agent-specific retrieval setups, this approach does not scale in enterprises, as it gives rise to knowledge silos and rule duplicates, and makes consistent updates and learning across agents difficult. We argue that this calls for an organizational memory for agentic business process execution: a shared, governed, and agent-consumable reference layer of evolving organization-specific procedural knowledge about how work should be executed. We derive requirements for such a memory, propose an architecture for its curation and consumption, and demonstrate its effectiveness in a proof-of-concept based on a procurement scenario.  \nKeywords: Agentic Process Automation · Large Language Models · Memory · Business Process Management.  \n1 Introduction  \nOrganizations are increasingly adopting LLM-based agents to execute business processes [4,7] . Unlike deterministic automation approaches such as robotic process automation (RPA), which are confined to structured and repetitive settings, these agents can interpret natural language and autonomously handle multistep tasks including exception handling. Thus, they offer a path to automating processes that were previously difficult to formalize with rule-based systems [26] .  \nHowever, general-purpose LLMs lack access to organization-specific knowledge by default. They are not inherently aware of company policies, process conventions, role structures, system landscapes, or exception-handling practices. Yet effective process execution in organizational settings depends precisely on such contextual knowledge, which varies across organizations, departments, and individual teams. This knowledge gap poses a fundamental challenge to scaling agentic business process execution in enterprise settings.  \nA natural response is to encode organization-specific knowledge in prompts or retrieval setups specific to individual agents [17, 16] . However, this approach does  \n2 L. Kirchdorfer et al.  \nnot scale: in large organizations, thousands of agents are expected to operate simultaneously across evolving processes and heterogeneous knowledge sources. Per-agent knowledge encoding creates new silos, where rules are duplicated, updated inconsistently, and interpreted differently across agents. At the sametime, organizations already possess vast amounts of process-relevant knowledge, represented in BPMN models, standard operating procedures (SOPs), policies, knowledge bases, and process analyses based on historical execution data. This knowledge is fragmented across artifacts, many of which were designed for human interpretation, not for LLM agent consumption. The core challenge is, therefore, not knowledge availability, but its transformation into a unified, agent-consumable form and its consistent provision to all agents in the organization.  \nWe argue that what is needed is an organizational memory for agentic business process execution: a shared, human-governed, and agent-consumable reference layer of organization-specific procedural knowledge about how work should be executed. Such a memory must integrate distributed knowledge from heterogeneous sources into a unified representation that agents can consume at runtime. It must provide the right context to the right agent","cbCaimaqLPrGyQPa","https://ap.wps.com/l/cbCaimaqLPrGyQPa","pdf",3883782,4,1,16,"English","en",105,"# Introduction\n# Background and Related Work\n## Agentic Business Process Management","[{\"question\":\"Why do general-purpose LLMs struggle with agentic business process execution in enterprises?\",\"answer\":\"They are not inherently aware of organization-specific context such as company policies, process conventions, role structures, system landscapes, and exception-handling practices, which are essential for reliable execution.\"},{\"question\":\"What problem arises when knowledge is encoded separately in per-agent prompts or retrieval setups?\",\"answer\":\"In large organizations, thousands of agents operate across evolving processes and heterogeneous sources; per-agent encoding leads to knowledge silos with duplicated and inconsistently updated rules, making consistent learning and updates difficult.\"},{\"question\":\"What is the proposed solution and how is it validated?\",\"answer\":\"The document proposes an organizational memory: a shared, governed, agent-consumable reference layer that integrates and evolves procedural process knowledge. It derives requirements, presents an architecture for curation and consumption, and demonstrates effectiveness using a procurement proof-of-concept.\"}]",1784189865,40,{"code":4,"msg":31,"data":32},"ok",{"site_id":25,"language":24,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"organizational-memory-for-agentic-business-process-execution","",{"@graph":36,"@context":85},[37,53,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":20},"https://docshare.wps.com/document/organizational-memory-for-agentic-business-process-execution/83705/",{"url":52,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":24,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":41,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-26","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"Why do general-purpose LLMs struggle with agentic business process execution in enterprises?","Question",{"text":75,"@type":76},"They are not inherently aware of organization-specific context such as company policies, process conventions, role structures, system landscapes, and exception-handling practices, which are essential for reliable execution.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What problem arises when knowledge is encoded separately in per-agent prompts or retrieval setups?",{"text":80,"@type":76},"In large organizations, thousands of agents operate across evolving processes and heterogeneous sources; per-agent encoding leads to knowledge silos with duplicated and inconsistently updated rules, making consistent learning and updates difficult.",{"name":82,"@type":73,"acceptedAnswer":83},"What is the proposed solution and how is it validated?",{"text":84,"@type":76},"The document proposes an organizational memory: a shared, governed, agent-consumable reference layer that integrates and evolves procedural process knowledge. It derives requirements, presents an architecture for curation and consumption, and demonstrates effectiveness using a procurement proof-of-concept.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,119,122,127,130,134],{"id":21,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":20,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":29,"slug":118},7,"Healthcare","healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":120,"slug":121},30,"research-report",{"id":123,"doc_module":4,"doc_module_name":46,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":46,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]