【新书速递】Scenarios Engineering for Services Science

Towards Trustworthy and Reliable Service Systems with New AI

Xuan Li, Fei-Yue Wang

Springer-Nature, 2026

The evolution of services science has been fundamentally reshaped by the convergence of Artificial Intelligence (AI), deep learning, and Large Language Models (LLMs). Services science-once centered on human-driven design and empirical business intuition-now faces a new era in which intelligent agents, digital ecosystems, and autonomous intelligence are becoming indispensable actors in value co-creation. Yet, this transformation is not merely technological: it demands a new philosophy for how service systems are conceived, verified, and governed. This brief presents Scenarios Engineering for Services Science (SESS) as a paradigmatic framework to safeguard the development of future service systems, including intelligent finance, intelligent customer services, intelligent hotels, intelligent cities, intelligent logistics, intelligent healthcare, and intelligent vehicles.

Rather than relying on brittle feature selection or intuition-based workflow design, SESS constructs service intelligence through carefully structured scenarios. These representations capture the evolving semantics of service activities-real or synthetic, physical or digital, individual or societal-and couple them with systematic mechanisms of Intelligence & Index, Calibration & Certification, and Verification & Validation. Based on ACP (Artificial societies, Computational experiments, and Parallel execution), SESS aims to liberate services science from isolated models and opaque heuristics, transforming it into calibrated, verifiable and reliables.

Key highlights include:

Scenarios Engineering for Services Science

Services science is ultimately relational: customers, providers, cyber infrastructures, and social constraints co-evolve within bounded spaces and times. Scenarios are the connective tissue of this ecosystem. They operate as the semantic substrate upon which new Al components-foundation models, autonomous agents, generative sim-ulators, and decentralized intelligence-can be trained, evaluated, and deployed. In addition, SESS asks a deeper question: "Under which scenarios can intelligent service act efficiently, effectively, safely, sustainably, and reliably?" In this regard, scenarios engineering do for new Al what experimental frameworks did for traditional engineering-they turn uncertain complexity into controllable, measurable, and certifiable intelligence.

Parallel Service Systems by Parallel Intelligence

Traditional Al has excelled in prediction but often failed under rare, extreme, or poorly calibrated service conditions. Inspired by Parallel Intelligence, SE4SS introduces a service-centric methodology. Through this integration, SEASS dissolves the boundaries between laboratory models, operational practices, and regulatory com-pliance. It shifts services from reactive management to proactive, scenario-aware orchestration, enabling autono To ensure intelligent service, we advocate guiding service systems toward a "6S" paradigm: service systems that are Safe in the physical world, Secure in the cyberspace, Sustainable in the ecological development, Sensitive to privacy, individual rights, and responsible resource utilization, capable of providing Service to All, and fostering the Smartness of All.

Future Service Systems with New Al

We foresee that future service systems, under the paradigm of parallel worlds, will operate in three modes. For the vast majority of time, service systems will exist in Autonomous Modes (AM), governed by swarms of digital agents that coordinate more than 80% of service operations. Parallel Modes (PM) will be employed for concrete execution, where autonomous service robots carry out critical tasks, with their presence limited to less than 15%, acting as the system's mechanical guardians. Expert/Emergency Mode (EM) is activated only under extreme events. In such cases, service scientists perform critical tasks on-site, supported by digital agents and autonomous service robots, and their involvement accounts for no more than 5% of all interactions.

While this brief aspires to advance the theoretical, methodological, and practical foundations of SESS, we remain acutely aware that this work represents only an initial step toward a more systematic understanding of service intelligence. We sincerely welcome comments, corrections, and suggestions from colleagues, students, and readers from all related communities. Any valuable insights-whether theoretical refinements, empirical evidence, or conceptual disagreements-will serve as a meaningful contribution to the continued development of this research.

November 24, 2025                                                                                                                                       Xuan Li

                                                                                                                                                                       Fei-Yue Wang