Official account of the DataFun community, dedicated to sharing big data and AI industry summit news and speaker talks, with regular downloadable resource packs.
Six leading Chinese companies — JD.com, Hangzhou Qunhe, Guanyuan Data, Datastrato, Li Auto, and AWS — share their production-grade context engineering practices for AI agents, covering layered memory, semantic layers, evidence chains, and action loops at DACon 2026.
Knora 4.2 introduces an AI-FDE closed loop that automates ontology construction, knowledge extraction, skill building, and agent execution, enabling proactive AI across five enterprise scenarios—production quality, operations tracing, professional analysis, R&D testing, and cost management—with measured results like 87.5% faster defect investigation and 80% reduction in report writing time.
XTransfer achieved 98.5% automated transaction review and 0.003% fraud rate in B2B cross-border trade by building a three-layer AI system: a B2B trade knowledge graph for data standardization, a vertical TradePilot model for document extraction and fraud detection, and an Agent assurance framework with sandbox simulation, A/B testing, and explainable AI for production deployment.
Palantir's August 2026 update to AIP Analyst introduces Skills (reusable analysis instructions) and Analysis Lookup (historical analyses as templates), adding a method-reuse layer to agent memory beyond knowledge retrieval, tightly integrated with Ontology for governed enterprise AI analysis.
OpenAI research shows that compaction summaries in long-horizon agents can inject malicious constraints or error-handling strategies into future context windows, creating a new state injection risk where model-generated intermediate states persist and corrupt subsequent agent behavior across multiple context boundaries.
Palantir CEO Alex Karp argues that AI splits enterprises into those with domain-specific AI-enhanced infrastructure and those without, warning that generic AI tools create no moat; real advantage comes from embedding AI into unique tribal knowledge, validated in extreme scenarios like Project Maven, and transferred to commercial clients through ontology-driven customization.
TypeSafe AI's Jev introduces a specialized 'System One Model' that replaces free-form LLM generation with fast, cheap structured decisions (Choice, Score, Noul) for high-frequency Agent control tasks like tool routing, guardrails, and context compaction, revealing an emerging Decision Layer in Agent architecture.
Amap shares four conference talks detailing how they built trustworthy AI agents for business decisions, covering semantic governance, causal evaluation, evaluation-driven development, and a product architecture shift that lifted accuracy from 50% to 95% by separating deterministic computation from model reasoning.
This article analyzes how 10 leading financial institutions implement AI agents in low-tolerance scenarios, detailing their approaches to data ontology, semantic layers, multi-agent architectures, risk control, and evaluation frameworks, achieving metrics like 98.5% automated review rates and 0.003% fraud rates while ensuring auditability and regulatory compliance.
This article explores how ontology-driven architecture solves the uncontrollability of LLM agents in enterprise settings by replacing external prompt-based constraints with explicit semantic modeling, detailing Knora's three-layer system that enables deterministic verification, precise context retrieval, and continuous ontology evolution from execution feedback, with real-world 70x efficiency gains in railway reporting.
