An Organizational Theory for Multi-Agent Interactions: Bridging Human Agents, LLMs, and Specialized AI

An Organizational Theory for Multi-Agent Interactions: Bridging Human Agents, LLMs, and Specialized AI | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article An Organizational Theory for Multi-Agent Interactions: Bridging Human Agents, LLMs, and Specialized AI Uwe M. Borghoff, Paolo Bottoni, Remo Pareschi This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6566773/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 02 Jul, 2025 Read the published version in Discover Computing → Version 1 posted You are reading this latest preprint version Abstract Purpose: Recent advances in AI, especially in large language models (LLMs), have created new opportunities to integrate human and artificial agents through shared linguistic capabilities. This paper presents a multi-agent organizational framework in which human agents, LLMs, and specialized agents (narrow AIs) collaborate via dynamic, topic-based group formation. Topic-driven interactions enable agents to coalesce around evolving interests, supported by threshold-based protocols for temporal adaptation, topic emergence, and participation. Methods: Within our framework, human agents guide the overall system objectives, while consultant agents (LLMs) provide semantic analysis and mediation, and specialized agents perform focused domain tasks. By leveraging automated topic modeling, the approach eschews rigid ontologies and instead supports adaptive and interpretable content management. Mathematical properties ensure system coherence—across roles, tasks, and timescales—while allowing natural evolution of interests and groups. Results: We illustrate the framework’s versatility with example scenarios in emergency response, healthcare research and financial decision-making, emphasizing how human decision-makers, LLM-based consultants, and specialized worker agents jointly fulfill complex goals through transparent topic alignment and threshold-driven coordination. This formalization advances human-computer interaction as a multi-agent phenomenon that integrates human insight with the strengths of next-generation AI models in a cohesive, evolving system. Artificial Intelligence and Machine Learning Theoretical Computer Science Human-Computer Interaction Large Language Models Topic-Based Group Formation Multi-Agent Systems Natural Language Processing Adaptive Interaction Protocols Full Text Additional Declarations The authors declare no competing interests. Cite Share Download PDF Status: Published Journal Publication published 02 Jul, 2025 Read the published version in Discover Computing → Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6566773","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":450514979,"identity":"44383a9f-45b3-4ba5-b8c2-eaff67c2b7c7","order_by":0,"name":"Uwe M. Borghoff","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA50lEQVRIiWNgGAWjYJACZhBhwMD44MMHBgYeAyBHgkgtzIYzZ5CsZTYPmEFAi8GN5APMBRX37LazNzM227bdkTGXSGC88QGvlrQE5hlnipN39hxmbM5te8ZjOSOB2XIGPi23cwyYedsSkg1u5B9/nNt2mMfgdgKbNA9eLfkfoFqSGZstYVr+4LeFAaTFDqyFEaYFn/cl7z8zODzjTEKCwZnDjI095w7zWM5/2GzZg0cL35nDDx8XVCTYGxxvZmz4UXbY3pzn8MEbP/BZAwQHgDixAcFnbMCuDg3YE6VqFIyCUTAKRiYAAOpNUaC0QYAWAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0002-7688-2367","institution":"University of the Bundeswehr Munich","correspondingAuthor":true,"prefix":"","firstName":"Uwe","middleName":"M.","lastName":"Borghoff","suffix":""},{"id":450514980,"identity":"f6debe1e-c133-4876-8292-c6c31b88f582","order_by":1,"name":"Paolo Bottoni","email":"","orcid":"https://orcid.org/0000-0003-4662-2019","institution":"Sapienza University of Rome","correspondingAuthor":false,"prefix":"","firstName":"Paolo","middleName":"","lastName":"Bottoni","suffix":""},{"id":450514981,"identity":"b88d2c0a-ab60-4c6d-9918-092f127ff178","order_by":2,"name":"Remo Pareschi","email":"","orcid":"https://orcid.org/0000-0002-4912-582X","institution":"University of Molise","correspondingAuthor":false,"prefix":"","firstName":"Remo","middleName":"","lastName":"Pareschi","suffix":""}],"badges":[],"createdAt":"2025-04-30 16:42:51","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-6566773/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6566773/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s10791-025-09667-2","type":"published","date":"2025-07-03T00:00:00+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":87759078,"identity":"91220105-7832-42fb-b146-8e18b0cdc6c8","added_by":"auto","created_at":"2025-07-28 16:25:53","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":788788,"visible":true,"origin":"","legend":"","description":"","filename":"2025RESEARCHSQUAREAnOrganizationalTheoryforMultiAgent.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6566773/v1_covered_ab21b6e6-ce92-40be-b3d4-3d070d5e700b.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003eAn Organizational Theory for Multi-Agent Interactions: Bridging Human Agents, LLMs, and Specialized AI\u003c/p\u003e","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Human-Computer Interaction, Large Language Models, Topic-Based Group Formation, Multi-Agent Systems, Natural Language Processing, Adaptive Interaction Protocols","lastPublishedDoi":"10.21203/rs.3.rs-6566773/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6566773/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003ePurpose: Recent advances in AI, especially in large language models (LLMs), have created new opportunities to integrate human and artificial agents through shared linguistic capabilities. This paper presents a multi-agent organizational framework in which human agents, LLMs, and specialized agents (narrow AIs) collaborate via dynamic, topic-based group formation. Topic-driven interactions enable agents to coalesce around evolving interests, supported by threshold-based protocols for temporal adaptation, topic emergence, and participation.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eMethods: Within our framework, human agents guide the overall system objectives, while consultant agents (LLMs) provide semantic analysis and mediation, and specialized agents perform focused domain tasks. By leveraging automated topic modeling, the approach eschews rigid ontologies and instead supports adaptive and interpretable content management. Mathematical properties ensure system coherence—across roles, tasks, and timescales—while allowing natural evolution of interests and groups.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eResults: We illustrate the framework’s versatility with example scenarios in emergency response, healthcare research and financial decision-making, emphasizing how human decision-makers, LLM-based consultants, and specialized worker agents jointly fulfill complex goals through transparent topic alignment and threshold-driven coordination. This formalization advances human-computer interaction as a multi-agent phenomenon that integrates human insight with the strengths of next-generation AI models in a cohesive, evolving system.\u003c/p\u003e","manuscriptTitle":"An Organizational Theory for Multi-Agent Interactions: Bridging Human Agents, LLMs, and Specialized AI","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-05-02 13:52:17","doi":"10.21203/rs.3.rs-6566773/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"6276824c-d5c0-4fe2-b0ed-c1f6a8cdd387","owner":[],"postedDate":"May 2nd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":47937316,"name":"Artificial Intelligence and Machine Learning"},{"id":47937317,"name":"Theoretical Computer Science"}],"tags":[],"updatedAt":"2025-07-28T16:25:47+00:00","versionOfRecord":{"articleIdentity":"rs-6566773","link":"https://doi.org/10.1007/s10791-025-09667-2","journal":{"identity":"discover-computing","isVorOnly":false,"title":"Discover Computing"},"publishedOn":"2025-07-03 00:00:00","publishedOnDateReadable":"July 3rd, 2025"},"versionCreatedAt":"2025-05-02 13:52:17","video":"","vorDoi":"10.1007/s10791-025-09667-2","vorDoiUrl":"https://doi.org/10.1007/s10791-025-09667-2","workflowStages":[]},"version":"v1","identity":"rs-6566773","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6566773","identity":"rs-6566773","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}