An engineering case study of a production low-code workflow on the n8n platform that integrates twelve RSS feeds from five patent offices that identifies five design lessons: P1 (Hybrid Intelligence Architecture), P2 (Format Normalization at Boundaries), P3 (Separation of Content and Presentation), P4 (Graceful Degradation), and P5 (Configuration Externalization), each supported by differentiated within-case evidence.
Abstract
Patent intelligence is hard to automate: data is heterogeneous and multilingual, and monitoring runs on daily cycles. This paper reports an engineering case study of a production low-code workflow on the n8n platform that integrates twelve RSS feeds from five patent offices (EPO, WIPO, USPTO, TIPO, and MOIP). Over a four-month deployment (October 2025 to January 2026; an approximately 120-day window), the system processed 340 items, normalized six timestamp formats, handled English, Chinese, and Korean, and generated three stakeholder-specific output formats with no manual intervention in formatting or delivery. From this deployment, we identify five design lessons: P1 (Hybrid Intelligence Architecture), P2 (Format Normalization at Boundaries), P3 (Separation of Content and Presentation), P4 (Graceful Degradation), and P5 (Configuration Externalization), each supported by differentiated within-case evidence. Output quality was assessed exploratorily with two LLM-based evaluators, whose inter-rater agreement was low for semantic dimensions; a small expert pilot (four English-language items, four raters) provided only a preliminary reference, on which the system output did not exhibit any obvious serious errors. Applying the lessons in other domains is future work, not a contribution. The findings are documented engineering experience from a single production case.
Organizations produce significant amounts of first-party data as part of their day-to-day operations in manufacturing, clinical, and customer experience domains, which are not accessible to generic large language models. The paper discusses the five-layer approach to using this information asset: (1) data ingestion and quality assurance, (2) domain ontology engineering, (3) knowledge graph engineering and population, (4) GraphRAG-enabled AI augmentation, and (5) downstream application enablement. The paper focuses on the design considerations, implementation tactics, and lessons learned from real-world applications in manufacturing, healthcare, and professional networks rather than presenting original research results. The implementation results show up to 70–80% query time reduction and close to 85% fewer hallucinations on average for GraphRAG over the standard RAG for most of the use cases analyzed. The main barriers to adoption are the substantial manual effort involved in ontology engineering, 73–94% entity resolution accuracy across different industries, and 3–5× higher computational costs for GraphRAG compared to RAG. The framework gives practitioners and researchers a reference architecture for designing, evaluating, and governing enterprise knowledge graph deployments built on proprietary organizational data.
Patent databases represent one of the largest public archives of technical knowledge, yet much of this knowledge remains difficult to identify, interpret, and reuse once patent rights expire or lapse. This paper proposes an AI-enabled framework for discovering expired and lapsing patents, identifying technology trends, and translating patent disclosures into business pathways. We use pathways to mean structured commercialization routes such as SaaS products, services, licensing packages, consulting playbooks, training offerings, data products, or internal process tools. The framework treats patent expiry as both a business signal and an archival transition, not primarily as a legal problem. Legal status remains important, but it is one risk-screening input alongside customer need, implementation feasibility, channel access, and market timing. We describe a system architecture that combines patent metadata, maintenance-fee records, legal-status indicators, semantic search, patent-family analysis, market signals, and generative AI workflows. A proof of concept parses all 378 records in an official weekly CIPO ST.96 archive, identifies 20 expired, lapsed, or near-expiry candidates, tests the stability of the transparent scoring model, and uses a locally hosted Qwen3.6 model to populate structured review packets. The evaluation demonstrates reproducible ingestion, stable rankings under weight perturbation, and schema-conformant model output, while also exposing incomplete legal-status coverage and the need for register and expert review. We argue that AI can function as a discovery and translation layer for dormant technical knowledge, but that such systems must explicitly represent legal uncertainty, data limitations, and commercialization risk.
Decades-old enterprise-resource-planning (ERP) systems lock operational data inside unstructured, human-readable reports, forcing slow, costly, error-prone manual re-keying. Because multimodal large language model (MLLM) capability is uneven, deploying MLLMs for extraction means trusting outputs without a labeled reference. We test this with a within-document controlled experiment on 400 controlled-substance stock-ledger documents (2951 records, 11 fields, predominantly Thai) from a Thai pharmaceutical factory, comparing trained human double-entry against four open-weight MLLMs (2 × 2 design: vendor × architecture) via OpenRouter. Human double-entry left 14 discrepancies against the adjudicated gold standard, none common to both operators. The strongest model, Qwen3-VL-32B-Instruct (Dense), reached 93.95% cell accuracy; among these four models, field accuracy varied more across vendors, whereas structural completeness differed consistently between dense models (0 missing records) and Mixture-of-Experts models (up to 51 of 2951 dropped). Deterministic accounting invariants flagged 0.61% of its records, leaving the unflagged majority 94.1% accurate across all 11 fields; adding calendar rules flagged 4.61% and raised residual date accuracy from 92.1% to 95.9%. We report both operating points and recommend the extended level where date fidelity is regulatory-critical. The pipeline is 13.5–29.4× faster in wall-clock terms and 97.5–99.5% cheaper. Gold-free, rule-based verification thus locates where MLLM reliability holds, giving human–AI collaboration quantified, disclosed residual risk rather than an implied guarantee. Even at the more conservative operating point, unflagged records average 94.5% accuracy across all 11 fields but only 43.7% on the free-text Remarks field, which the triage cannot check; the results support risk reduction and the localization of review effort, not unrestricted regulatory reliability across all fields.
Enterprise compliance management requires rapid adaptation to evolving regulatory frameworks (e.g., DORA, AI RMF, FedRAMP) and tight remediation SLAs. Traditional static orchestrators often fail in hybrid cloud environments where event-driven assessments demand that automation code adapt to runtime context in seconds. This paper presents lessons learned from evaluating six large language models for AI-driven workflow generation in a production enterprise platform, benchmarked across 29 real-world IT automation scenarios, two generation pipeline architectures, and eight independent runs per prompt-model-pipeline configuration (2,784 runs total). Our initial pipeline used monolithic workflow generation, achieving 31.5-82.8% structural success rates (JSON schema validity and correct UI rendering), with most models struggling on complex JSON generation. We developed a redesigned piecewise pipeline that decomposes workflow construction into variable scaffolding, base block assembly, and nested block generation, raising structural success to 74.1-97.8% across all models. We analyze production tradeoffs including cost (USD 0.008-0.20 per workflow), latency (under 50s for interactive use), and model selection. Piecewise decomposition enables smaller models (e.g., mistral-small at 95.7% structural success and USD 0.01 per workflow) to reach production viability, removing dependency on expensive frontier models. While mistral-medium-2505 and gpt-oss-120b achieved the highest structural success (96.1% and 97.8%), mistral-medium-2505 carries a 19x cost premium versus mistral-small. Our deployment lessons highlight the need to separate structural validity from semantic correctness (logical fulfillment of user intent) and provide a solution for model-agnostic, scalable automation in cloud engineering.
Xavier Wrenn, Radoslav Raykov, Aleksandar Angelov et al.· 0 citations
Penetration testing reports are a critical artifact in the cybersecurity workflow, yet their technical complexity frequently limits their utility for non-specialist stakeholders involved in risk remediation decisions. This paper investigates the feasibility of using four open-weight large language models, DeepSeek-r1:32b, Qwen3.5:35b, Gemma4:31b, and GLM-4.7-flash:32b, to generate plain-language summaries of penetration testing reports. A corpus of 65 publicly available pentest reports was used for evaluation. Model outputs were assessed across four dimensions: readability, technical term density, semantic similarity to the source document, and factual correctness using LLM-as-a-judge evaluation. Two classical extractive methods, LSA and TextRank, were included as baselines. Readability analysis using seven established metrics showed that Qwen3.5 and Gemma4 produced the most accessible summaries, reducing mean Flesch Reading Ease scores from 25.2 in the originals to 49.2 and 52.1 respectively, and lowering grade-level scores from post-graduate to high-school equivalents. Results across the remaining evaluation dimensions further indicate that appropriately selected open-weight LLMs can produce accessible and factually grounded summaries of technical security documents, offering a practical alternative to proprietary solutions in privacy-sensitive deployment contexts.
Prerit Datta, M. Islam, Ryan Wojciechowski· Annual International Compute...· 0 citations
The operationalization of machine learning (ML) introduces distinct engineering and lifecycle management challenges—such as extreme data dependence, silent model degradation (concept drift), and inherent non-determinism—which traditional software engineering workflows fail to adequately address. This systematic literature review provides a rigorous, comprehensive mapping of the ML lifecycle domain between 2015 and 2025 using the PRISMA protocol. Out of an initial pool of 12,450 articles, a highly specialized cohort of 22 primary studies was extracted, classified, and synthesized to map out contemporary Machine Learning Operations (MLOps) patterns, technical debt structures, governance models, and security vulnerabilities. To address the documented “production gap,” this paper formalizes the findings into a synthesized operational mapping and introduces a preliminary conceptual layout for an Adaptive Lifecycle Framework (ALF), juxtaposing it with legacy paradigms like CRISP-DM. Furthermore, we expand the scope to investigate domain-specific lifecycle complexities in healthcare systems and Large Language Model (LLM) pipelines, providing an essential evolutionary baseline for sustainable MLOps.
Ioannis-John Kosmas, Theofanis Papadopoulos, C. Michalakelis· AppliedMath· 0 citations
