Do Enterprise Systems Need Learned World Models? The Importance of Context to Infer Dynamics

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Nair, Jishnu Sethumadhavan, Bechard, Patrice, Maheshwary, Rishabh, Dasgupta, Surajit, Ramachandran, Sravan, Bhagat, Aakash, Radhakrishna, Shruthan, Pattnaik, Pulkit, Obando-Ceron, Johan, Malay, Shiva Krishna Reddy, Davasam, Sagar, Subramanian, Seganrasan, Mittal, Vipul, Nemala, Sridhar Krishna, Pal, Christopher, Sunkara, Srinivas, Rajeswar, Sai
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916005345755136
author Nair, Jishnu Sethumadhavan
Bechard, Patrice
Maheshwary, Rishabh
Dasgupta, Surajit
Ramachandran, Sravan
Bhagat, Aakash
Radhakrishna, Shruthan
Pattnaik, Pulkit
Obando-Ceron, Johan
Malay, Shiva Krishna Reddy
Davasam, Sagar
Subramanian, Seganrasan
Mittal, Vipul
Nemala, Sridhar Krishna
Pal, Christopher
Sunkara, Srinivas
Rajeswar, Sai
author_facet Nair, Jishnu Sethumadhavan
Bechard, Patrice
Maheshwary, Rishabh
Dasgupta, Surajit
Ramachandran, Sravan
Bhagat, Aakash
Radhakrishna, Shruthan
Pattnaik, Pulkit
Obando-Ceron, Johan
Malay, Shiva Krishna Reddy
Davasam, Sagar
Subramanian, Seganrasan
Mittal, Vipul
Nemala, Sridhar Krishna
Pal, Christopher
Sunkara, Srinivas
Rajeswar, Sai
contents World models enable agents to anticipate the effects of their actions by internalizing environment dynamics. In enterprise systems, however, these dynamics are often defined by tenant-specific business logic that varies across deployments and evolves over time, making models trained on historical transitions brittle under deployment shift. We ask a question the world-models literature has not addressed: when the rules can be read at inference time, does an agent still need to learn them? We argue, and demonstrate empirically, that in settings where transition dynamics are configurable and readable, runtime discovery complements offline training by grounding predictions in the active system instance. We propose enterprise discovery agents, which recover relevant transition dynamics at runtime by reading the system's configuration rather than relying solely on internalized representations. We introduce CascadeBench, a reasoning-focused benchmark for enterprise cascade prediction that adopts the evaluation methodology of World of Workflows on diverse synthetic environments, and use it together with deployment-shift evaluation to show that offline-trained world models can perform well in-distribution but degrade as dynamics change, whereas discovery-based agents are more robust under shift by grounding their predictions in the current instance. Our findings suggest that, in configurable enterprise environments, agents should not rely solely on fixed internalized dynamics, but should incorporate mechanisms for discovering relevant transition logic at runtime.
format Preprint
id arxiv_https___arxiv_org_abs_2605_12178
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Do Enterprise Systems Need Learned World Models? The Importance of Context to Infer Dynamics
Nair, Jishnu Sethumadhavan
Bechard, Patrice
Maheshwary, Rishabh
Dasgupta, Surajit
Ramachandran, Sravan
Bhagat, Aakash
Radhakrishna, Shruthan
Pattnaik, Pulkit
Obando-Ceron, Johan
Malay, Shiva Krishna Reddy
Davasam, Sagar
Subramanian, Seganrasan
Mittal, Vipul
Nemala, Sridhar Krishna
Pal, Christopher
Sunkara, Srinivas
Rajeswar, Sai
Artificial Intelligence
Computation and Language
Machine Learning
World models enable agents to anticipate the effects of their actions by internalizing environment dynamics. In enterprise systems, however, these dynamics are often defined by tenant-specific business logic that varies across deployments and evolves over time, making models trained on historical transitions brittle under deployment shift. We ask a question the world-models literature has not addressed: when the rules can be read at inference time, does an agent still need to learn them? We argue, and demonstrate empirically, that in settings where transition dynamics are configurable and readable, runtime discovery complements offline training by grounding predictions in the active system instance. We propose enterprise discovery agents, which recover relevant transition dynamics at runtime by reading the system's configuration rather than relying solely on internalized representations. We introduce CascadeBench, a reasoning-focused benchmark for enterprise cascade prediction that adopts the evaluation methodology of World of Workflows on diverse synthetic environments, and use it together with deployment-shift evaluation to show that offline-trained world models can perform well in-distribution but degrade as dynamics change, whereas discovery-based agents are more robust under shift by grounding their predictions in the current instance. Our findings suggest that, in configurable enterprise environments, agents should not rely solely on fixed internalized dynamics, but should incorporate mechanisms for discovering relevant transition logic at runtime.
title Do Enterprise Systems Need Learned World Models? The Importance of Context to Infer Dynamics
topic Artificial Intelligence
Computation and Language
Machine Learning
url https://arxiv.org/abs/2605.12178