Progressive Crystallization: Turning Agent Exploration into Deterministic, Lower-Cost Workflows in Production

Jul 16, 2026

Upload by Eden

AI agents for IT operations often act as permanent cost centers because each run re-invokes full LLM inference, even for recurring issues already solved before. Progressive crystallization reframes exploration as discovery: repeatedly validated agent behaviors are promoted into deterministic, zero-token workflows, with automatic demotion if regressions occur. A taxonomy links orchestration, hybrid execution, and fully deterministic playbooks. In a production cloud network agentic system, deterministic execution rose to 45%, per-incident agent cost dropped over 70%, incident volume doubled, and safety improved monotonically through promotions.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages4

Reading Time10 min

Introduction

Related Work

Execution-Type Taxonomy

Progressive Crystallization Lifecycle

Economic Model

Safety Monotonicity and Production Evidence

Limitations and Threats to Validity

Progressive Crystallization: Turning Agent Exploration into Deterministic, Lower-Cost Workflows in Production Arun Malik Microsoft Azure Networking Email: [arunma@microsoft.com](arunma@microsoft.com) ORCID: 0009-0005-6650-6711 arXiv :2607 .07052v 1 [ cs . SE] 8 Jul 2026 Abstract—AI agents deployed for IT operations are typically permanent cost centers: every execution re-invokes full LLM inference, even for problems the agent has already solved many times. We present progressive crystallization, a lifecycle that treats agent exploration as a discovery mechanism rather than a permanent execution model. We define an execution-type taxonomy spanning three points, from fully agent-orchestrated (stochastic, expensive) through hybrid to fully deterministic (zero-token, reproducible), and an evidence-based promotion mechanism that converts an agent’s repeatedly validated behavior into progressively cheaper and more deterministic workflows, with automatic demotion when a promoted workflow regresses. This makes anagentic platform cheaper, faster, and safer over time without human rewriting of agent-discovered patterns. In a production agentic system for cloud network operations handling tens of thousands of incidents per month, the share of executions served by deterministic workflows rose from zero to 45 percent over eight months, per-incident agent cost fell by more than 70 percent while incident volume doubled, and safety properties improved monotonically across promotions because determinism increases reproducibility and auditability. We describe the taxonomy, the promotion and demotion criteria, the trace-extraction method, an economic model, and the safety argument, and we discuss limitations and threats to validity. Index Terms—agentic AI, AIOps, workflow automation, LLM cost optimization, process mining, deterministic execution, progressive autonomy, safety I. INTRODUCTION Large language model (LLM) agents are increasingly deployed for IT operations tasks such as incident triage, rootcause analysis, and automated remediation [1], [2] . An agent observes system state through tool calls, reasons about the problem, and executes actions in a loop. This flexibility is exactly what lets an agent handle novel incidents, but it also makes the agent a permanent cost center. Every execution reinvokes full model inference, so the same recurring incident consumes thousands of tokens each time it recurs, the investigation path is non-deterministic and hard to reproduce, and total cost scales linearly with incident volume. The core waste is that a successful agent investigation is discarded. When an agent resolves an incident through a novel path, that knowledge is lost; the next occurrence of the same failure re-discovers the solution from scratch, at the same token cost and with a possibly different and inferior result. Existing options do not close this gap. Traditional workflow engines cannot handle novel scenarios and require an engineer- ing sprint per new automation. Unconstrained agents never get cheaper. Fine-tuning yields a smaller probabilistic model rather than a deterministic workflow, and recorded-macro or roboticprocess-automation approaches capture surface actions, not the reasoning or the data flow, and are brittle to environment change. We argue that agent exploration should be a discovery mechanism, not a permanent execution model. Behavior that an agent discovers and validates through repeated successful execution can be systematically converted into deterministic workflows that require zero LLM tokens to run, while the agent layer remains available for genuinely novel problems. We call this lifecycle progressive crystallization. This paper makes four contributions: • an execution-type taxonomy that places operational workflows on a spectrum from fully agent-orchestrated to fully deterministic (Section III); • a promotion and demotion lifecycle that advances workflows down the spectrum based

FAQs

Why do LLM agent deployments in IT operations become high-cost over time?

Each agent execution typically reruns full LLM inference, so recurring incidents consume thousands of tokens again. The investigation path is also non-deterministic, making results harder to reproduce and increasing end-to-end cost as volume grows.

What is progressive crystallization and how does it reduce cost?

Progressive crystallization treats agent exploration as discovery and promotes repeatedly validated behaviors into deterministic workflows. Promoted workflows require zero LLM tokens to run, while the agent remains available for truly novel problems.

How are promotion and demotion handled to maintain safety and performance?

The method uses evidence-based promotion criteria to move workflows down the spectrum toward determinism, and regression-based demotion to revert when promoted workflows degrade. The safety argument relies on monotonic improvement as determinism increases reproducibility and auditability.

Jul 16, 2026

Upload by Eden