# Four-Axis Decision Alignment for Long-Horizon Enterprise AI Agents

arXiv: https://arxiv.org/abs/2604.19457 · HTML: https://mlsanity.com/paper/2604.19457/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 19 of its 23 references that are in the map.
- First posted on arXiv 21 Apr 2026.

## Builds on (most cited first, 19 of 19)

- [Self-Consistency Improves Chain of Thought Reasoning in Language Models](https://mlsanity.com/paper/2203.11171/) (2022)
- [G-Eval: NLG Evaluation using GPT-4 with Better Human Alignment](https://mlsanity.com/paper/2303.16634/) (2023)
- [MemGPT: Towards LLMs as Operating Systems](https://mlsanity.com/paper/2310.08560/) (2023)
- [A-MEM: Agentic Memory for LLM Agents](https://mlsanity.com/paper/2502.12110/) (2025)
- [A General Language Assistant as a Laboratory for Alignment](https://mlsanity.com/paper/2112.00861/) (2021)
- [Evaluating Very Long-Term Conversational Memory of LLM Agents](https://mlsanity.com/paper/2402.17753/) (2024)
- [LongMemEval: Benchmarking Chat Assistants on Long-Term Interactive Memory](https://mlsanity.com/paper/2410.10813/) (2024)
- [Ragas: Automated Evaluation of Retrieval Augmented Generation](https://mlsanity.com/paper/2309.15217/) (2023)
- [Chain-of-Verification Reduces Hallucination in Large Language Models](https://mlsanity.com/paper/2309.11495/) (2023)
- [Evaluating Memory in LLM Agents via Incremental Multi-Turn Interactions](https://mlsanity.com/paper/2507.05257/) (2025)
- [MEM1: Learning to Synergize Memory and Reasoning for Efficient Long-Horizon Agents](https://mlsanity.com/paper/2506.15841/) (2025)
- [FLASK: Fine-grained Language Model Evaluation based on Alignment Skill Sets](https://mlsanity.com/paper/2307.10928/) (2023)
- [MIRIX: Multi-Agent Memory System for LLM-Based Agents](https://mlsanity.com/paper/2507.07957/) (2025)
- [Hierarchical Memory for High-Efficiency Long-Term Reasoning in LLM Agents](https://mlsanity.com/paper/2507.22925/) (2025)
- [AMA-Bench: Evaluating Long-Horizon Memory for Agentic Applications](https://mlsanity.com/paper/2602.22769/) (2026)
- [TiMem: Temporal-Hierarchical Memory Consolidation for Long-Horizon Conversational Agents](https://mlsanity.com/paper/2601.02845/) (2026)
- [GAM: Hierarchical Graph-based Agentic Memory for LLM Agents](https://mlsanity.com/paper/2604.12285/) (2026)
- [HyMem: Hybrid Memory Architecture with Dynamic Retrieval Scheduling](https://mlsanity.com/paper/2602.13933/) (2026)
- [Modality-Native Routing in Agent-to-Agent Networks: A Multimodal A2A Protocol Extension](https://mlsanity.com/paper/2604.12213/) (2026)

## Built on by (most cited first, 1 of 1)

- [Stateless Decision Memory for Enterprise AI Agents](https://mlsanity.com/paper/2604.20158/) (2026)

## Abstract (arXiv)

Long-horizon enterprise agents make high-stakes decisions (loan underwriting, claims adjudication, clinical review, prior authorization) under lossy memory, multi-step reasoning, and binding regulatory constraints. Current evaluation reports a single task-success scalar that conflates distinct failure modes and hides whether an agent is aligned with the standards its deployment environment requires. We propose that long-horizon decision behavior decomposes into four orthogonal alignment axes, each independently measurable and failable: factual precision (FRP), reasoning coherence (RCS), compliance reconstruction (CRR), and calibrated abstention (CAR). CRR is a novel regulatory-grounded axis; CAR is a measurement axis separating coverage from accuracy. We exercise the decomposition on a controlled benchmark (LongHorizon-Bench) covering loan qualification and insurance claims adjudication with deterministic ground-truth construction. Running six memory architectures, we find structure aggregate accuracy cannot see: retrieval collapses on factual precision; schema-anchored architectures pay a scaffolding tax; plain summarization under a fact-preservation prompt is a strong baseline on FRP, RCS, EDA, and CRR; and all six architectures commit on every case, exposing a decisional-alignment axis the field has not targeted. The decomposition also surfaced a pre-registered prediction of our own, that summarization would fail factual recall, which the data reversed at large magnitude, an axis-level reversal aggregate accuracy would have hidden. Institutional alignment (regulatory reconstruction) and decisional alignment (calibrated abstention) are under-represented in the alignment literature and become load-bearing once decisions leave the laboratory. The framework transfers to any regulated decisioning domain via two steps: build a fact schema, and calibrate the CRR auditor prompt.