{
  "abstract": "Enterprise information systems remain strongly organized around persistent entities, CRUD operations, and prede-\n\nfined workflow states. These mechanisms are effective for deterministic transactions, but provide limited architectural\n\nsupport for interactions whose desired outcome is expressed as an evolving goal over heterogeneous context. Large\n\nLanguage Model (LLM) agents provide adaptive planning and tool orchestration, yet their probabilistic execution\n\nintroduces challenges concerning bounded execution, authorization, state divergence, memory continuity, and au-\n\nditable control. This paper proposes the Autonomous Relationship Engine (ARE), a Design Science Research (DSR)\n\nartifact centered on a persistent, time-bounded Intent Node as a unit of enterprise agentic execution. An Intent Node\n\nis modeled as I(t) = <G, C, P, M(t), S(t)>, where Gt represents an operational goal specification, Ct hard constraints,\n\nPt runtime policy, Mt cognitive memory, and St authoritative execution-domain state. ARE distinguishes incoming\n\nintent from an active commitment and defines commitment-scoped action admissibility over the active commitment,\n\nauthoritative state, action scope, and deterministic policy decision. Goal satisfaction is evaluated through &Phi;(Gt , St ),\n\nwhile candidate actions are generated through probabilistic planning and subjected to a Policy Decision Point (PDP)\n\nexpressed through Policy-as-Code (PaC) before controlled execution. A tri-modal memory model provides episodic,\n\nsemantic, and procedural context while remaining non-authoritative for authorization. The paper formalizes the\n\ncommitment lifecycle, goal-state evaluation, bounded execution, policy enforcement, memory authority boundaries,\n\nand a reproducible experimental protocol, including a falsifiable ablation of commitment-linked cognitive-context re-\n\nconstruction after interruption, state change, policy update, and session resumption. The contribution is positioned\n\nas an architectural synthesis and formalization rather than a claim of novel individual mechanisms or experimentally\n\nestablished performance; the proposed evaluation criteria remain hypotheses for subsequent validation.",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": null,
  "arxiv": null,
  "author_total": 1,
  "authors": [
    {
      "name": "Edison Ramón Sira Carreño",
      "orcid": "0009-0005-5577-8906"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/bvqvg1ns1easd.bib",
  "cited_by": 0,
  "doi": "10.5281/zenodo.23489471",
  "fulltext_url": null,
  "id": "bvqvg1ns1easd",
  "issns": [],
  "landing_url": "https://zenodo.org/records/23489471",
  "language": "eng",
  "licence": "CC-BY-4.0",
  "licence_label": "CC BY 4.0",
  "licence_source": "article",
  "licence_url": "https://creativecommons.org/licenses/by/4.0/",
  "metadata_licence": null,
  "native_subjects": [],
  "openalex": null,
  "pdf_url": "https://zenodo.org/api/records/23489471/files/ARE_Preprint_ARXIV_FINAL_CLEAN.pdf/content",
  "pmcid": null,
  "pmid": null,
  "published": "2026-10-20",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/bvqvg1ns1easd.ris",
  "sources": [
    {
      "first_seen": "2026-11-03T12:20:21Z",
      "native_id": "23489471",
      "source": "zenodo",
      "url": "https://zenodo.org/records/23489471"
    }
  ],
  "subjects": [],
  "title": "Autonomous Relationship Engine: Persistent Intent, Goal-State Evaluation, and Cognitive Memory for Enterprise Agentic Execution",
  "title_text": "Autonomous Relationship Engine: Persistent Intent, Goal-State Evaluation, and Cognitive Memory for Enterprise Agentic Execution",
  "type": "preprint",
  "updated_at": "2026-11-03T12:20:21Z",
  "url": "https://research.draconitic.com/paper/bvqvg1ns1easd",
  "venue": "Zenodo",
  "versions": []
}