AI Insights

• The system is designed as a multi-agent architecture with 7 agents working together to provide supply chain risk management and mitigation strategies. [2]
• The system uses few-shot prompting to provide illustrative examples of inputs paired with structured outputs, demonstrating the desired format and reasoning style. [1]

Abstract
Modern supply chains are increasingly exposed to disruptions from geopolitical events, demand shocks, trade restrictions, to natural disasters. While many of these disruptions originate deep in the supply network, most companies still lack visibility beyond Tier-1 suppliers, leaving upstream vulnerabilities undetected until the impact cascades downstream. To overcome this blind-spot and move from reactive recovery to proactive resilience, we introduce a minimally supervised agentic AI framework that autonomously monitors, analyses, and responds to disruptions across extended supply networks. The architecture comprises seven specialised agents powered by large language models and deterministic tools that jointly detect disruption signals from unstructured news, map them to multi-tier supplier networks, evaluate exposure based on network structure, and recommend mitigations such as alternative sourcing options. \rev{We evaluate the framework across 30 synthesised scenarios covering three automotive manufacturers and five disruption classes. The system achieves high accuracy across core tasks, with F1 scores between 0.962 and 0.991, and performs full end-to-end analyses in a mean of 3.83 minutes at a cost of \$0.0836 per disruption. Relative to industry benchmarks of multi-day, analyst-driven assessments, this represents a reduction of more than three orders of magnitude in response time. A real-world case study of the 2022 Russia-Ukraine conflict further demonstrates operational applicability. This work establishes a foundational step toward building resilient, proactive, and autonomous supply chains capable of managing disruptions across deep-tier networks.

Why we are recommending this paper?
Due to your Interest in AI for Product Management

This paper directly addresses the challenges of product strategy and roadmap development by examining supply chain disruptions, a critical factor in product availability and market success. The agentic AI approach aligns with the user's interest in leveraging AI for proactive product management.

AI Insights

• The experiment results revealed essential factors that affect participants' online decision-making, including demographic factors like gender, contextual factors like domain-specific knowledge and behavioral factors like thinking time. [3]
• Domain-specific knowledge played a vital role in decision-making from two perspectives - (i) participants' contextual attitude regarding the topic (existing domain-specific knowledge) and (ii) information retrieval (new domain-specific knowledge intake). [3]
• The effect of AIGC differs under different circumstances. [3]
• AIGC can be as useful as human-written information, even when both are cited from appropriate sources, regardless of the user's domain-specific literacy. [3]
• AIGC: Artificial Intelligence-Generated Content RQ1: Research Question 1 RQ2: Research Question 2 The results demonstrate that AIGC can be as useful as human-written information, even when both are cited from appropriate sources, regardless of the user's domain-specific literacy. [3]
• Properly sourced AIGC is manageable since all information-related signals are transparent and thus subjects bear their own risk when making decisions. [2]
• Subjects tend to draw existing conclusions from the information rather than elaborate on the evidence or reasoning if they have insufficient professional knowledge. [1]

Abstract
Modelling users' online decision-making and opinion change is a complex issue that needs to consider users' personal determinants, the nature of the topic and the information retrieval activities. Furthermore, generative-AIbased products like ChatGPT gradually become an essential element for the retrieval of online information. However, the interaction between domainspecific knowledge and AI-generated content during online decision-making is unclear. We conducted a lab-based explanatory sequential study with university students to overcome this research gap. In the experiment, we surveyed participants about a set of general domain topics that are easy to grasp and another set of domain-specific topics that require adequate levels of chemical science knowledge to fully comprehend. We provided participants with decision-supporting information that was either produced using generative AI or collected from selected expert human-written sources to explore the role of AI-generated content compared to ordinary information during decision-making. Our result revealed that participants are less likely to change opinions on domain-specific topics. Since participants without professional knowledge had difficulty performing in-depth and independent reasoning based on the information, they favoured relying on conclusions presented in the provided materials and tended to stick to their initial opinion. Besides, information that is labelled as AI-generated is equivalently helpful as information labelled as dedicatedly human-written for participants in this experiment, indicating the vast potential as well as concerns for AI replacing human experts to help users tackle professional topics or issues.

Why we are recommending this paper?
Due to your Interest in AI for Product Management

Given the user's interest in AI for product management, this paper explores the influence of AI-generated content on decision-making, a key consideration for any product vision or strategy. Understanding how AI impacts user choices is crucial for effective product development.

AI Insights

• The pod coordination use case highlights significant challenges that quantum computing faces in practical applications, particularly in constraint-heavy optimization related vehicle routing problems. [3]
• Advancements in quantum hardware and algorithms may provide benefits in this area, but it is highly conditional on overcoming current limitations of qubit count and handling hard constraints and continuous variables effectively. [3]
• QUBO formulation: A mathematical representation of a problem using quadratic unconstrained binary optimization. [3]
• Quantum annealing: A quantum computing technique used to find the global minimum of a given objective function. [3]
• The use cases presented in this report demonstrate the challenges and limitations of applying quantum computing to real-world problems, particularly those related to constraint-heavy optimization and vehicle routing. [3]
• Advancements in quantum hardware and algorithms may provide benefits in these areas, but significant progress is needed to overcome current limitations. [3]
• Classical methods and heuristics have proven effective in solving these problems, making it essential for organizations to carefully evaluate the potential benefits of quantum computing before investing resources. [3]
• Pod coordination problem (PCP): A demand-responsive, door-to-door transport system utilizing passive modular containers (pods) carried by active vehicles such as trains or trucks. [2]
• Classical methods employed have proven effective and the existing heuristics outperform quantum approaches at this stage. [1]

Abstract
This whitepaper surveys the current landscape and short- to mid-term prospects for quantum-enabled optimization and machine learning use cases in industrial settings. Grounded in the QCHALLenge program, it synthesizes hardware trajectories from different quantum architectures and providers, and assesses their maturity and potential for real-world use cases under a standardized traffic-light evaluation framework. We provide a concise summary of relevant hardware roadmaps, distinguishing superconducting and ion-trap technologies, their current states, modalities, and projected scaling trajectories. The core of the presented work are the use case evaluations in the domains of optimization problems and machine learning applications. For the conducted experiments, we apply a consistent set of evaluation criteria (model formulation, scalability, solution quality, runtime, and transferability) which are assessed in a shared system of three categories, ranging from optimistic (solutions produced by quantum computers are competitive with classical methods and/or a clear path to a quantum advantage is shown) to pessimistic (significant hurdles prevent practical application of quantum solutions now and potentially in the future). The resulting verdicts illuminate where quantum approaches currently offer promise, where hybrid classical-quantum strategies are most viable, and where classical methods are expected to remain superior.

Why we are recommending this paper?
Due to your Interest in Product Strategy

Coming from BMW Group, this whitepaper offers strategic recommendations for quantum computing, a potentially transformative technology for future product development and innovation. This aligns with the user's interest in exploring emerging technologies and their impact on product strategy.

AI Insights

• Social Welfare Function: A function that captures economic efficiency by measuring the total value generated in the system rather than just monetary transfers. [3]
• The paper builds on the work of Little's Law, which relates the average number of customers in a system to the arrival rate and service rate. [3]
• The paper explores a multi-server queueing system with two products and derives the optimal inventory levels for each product that maximize social welfare. [2]

Abstract
This paper analyzes a two-product make-to-stock queueing system where a single production facility serves two customer classes with independent Poisson arrivals. Customers make strategic join-or-balk decisions without observing current inventory levels. The analysis establishes the existence and uniqueness of Nash equilibria in customer joining strategies for various inventory scenarios. Optimal base-stock levels are characterized from both profit-maximizing and welfare-maximizing perspectives, with closed-form expressions for key performance measures.

Why we are recommending this paper?
Due to your Interest in Product Management

This paper's focus on inventory management and strategic joining aligns with the user's interest in product roadmap development and optimizing product offerings. Understanding inventory levels is essential for efficient product strategy.

AI Insights

• Jailbreak attack: A type of adversarial attack that aims to manipulate a language model into producing harmful or undesirable outputs. [3]
• The paper discusses various methods for defending against jailbreak attacks on large language models. [2]
• The authors present a novel method called 'DiffusionAttacker' that uses diffusion-driven prompt manipulation to evade defenses. [1]

Abstract
Large Language Models (LLMs) have achieved remarkable capabilities but remain vulnerable to adversarial ``jailbreak'' attacks designed to bypass safety guardrails. Current safety alignment methods depend heavily on static external red teaming, utilizing fixed defense prompts or pre-collected adversarial datasets. This leads to a rigid defense that overfits known patterns and fails to generalize to novel, sophisticated threats. To address this critical limitation, we propose empowering the model to be its own red teamer, capable of achieving autonomous and evolving adversarial attacks. Specifically, we introduce Safety Self- Play (SSP), a system that utilizes a single LLM to act concurrently as both the Attacker (generating jailbreaks) and the Defender (refusing harmful requests) within a unified Reinforcement Learning (RL) loop, dynamically evolving attack strategies to uncover vulnerabilities while simultaneously strengthening defense mechanisms. To ensure the Defender effectively addresses critical safety issues during the self-play, we introduce an advanced Reflective Experience Replay Mechanism, which uses an experience pool accumulated throughout the process. The mechanism employs a Upper Confidence Bound (UCB) sampling strategy to focus on failure cases with low rewards, helping the model learn from past hard mistakes while balancing exploration and exploitation. Extensive experiments demonstrate that our SSP approach autonomously evolves robust defense capabilities, significantly outperforming baselines trained on static adversarial datasets and establishing a new benchmark for proactive safety alignment.

Why we are recommending this paper?
Due to your Interest in Vision Setting for Tech Teams

This research addresses safety alignment for large language models, a critical consideration for any AI-powered product management tool. The focus on adversarial attacks and mitigation strategies is directly relevant to ensuring responsible AI adoption within product teams.