Architectural Cases

The integration of agentic AI into business operations is a transformative trend, yet its true potential remains constrained by foundational data challenges. While enterprises are eager to embrace AI agents, many struggle to achieve a significant return on investment due to inadequate data infrastructure. The transition from AI merely answering questions to actively executing tasks demands seamless access to comprehensive, contextually rich data across all organizational systems. Traditional data architectures, even those recently updated, are often ill-equipped to meet these escalating demands, particularly for real-time decision-making and interaction with operational systems like supply chain or HR platforms.

A recent comprehensive study involving 300 data and technology leaders sheds light on the obstacles posed by outdated data systems to the effectiveness of AI agents. This research reveals a clear distinction between organizations that are successfully leveraging agentic AI and those that are not. A select group, identified as 'data leaders,' demonstrates superior outcomes by proactively addressing data limitations. Their success provides valuable insights and a practical roadmap for creating an optimal data ecosystem that fosters the growth and scalability of AI agents, thereby enabling more reliable and impactful AI deployments.

The Critical Role of Data Access in Agentic AI Success

The success of AI agents is profoundly influenced by their ability to access and utilize enterprise data. A recent study reveals that, on average, AI systems within surveyed organizations only have access to 45% of available company data. This figure drops significantly to 30% or less for organizations categorized as 'data laggards,' severely restricting the intelligence and effectiveness of their AI agents. In stark contrast, 'data leaders' – a small but impactful group – ensure their AI agents can access over 70% of their data, leading to noticeably superior performance and outcomes from their AI deployments. This disparity underscores that comprehensive data access is not just beneficial, but a prerequisite for maximizing the value of agentic AI.

Beyond mere access, the trustworthiness of decisions made by AI agents is directly correlated with the robustness of an organization's data foundation. Currently, only about half of the surveyed organizations express confidence in the accuracy and relevance of their AI agents' decisions. However, among 'data leaders,' this trust level reaches a resounding 100%, indicating a strong causal link between a reliable data infrastructure and confidence in AI-driven outputs. Furthermore, 'data leaders' find it considerably easier to scale their AI agent initiatives and achieve decision-making at speed. While two-thirds of 'data laggards' report that legacy data systems impede AI agent scaling (66%) and fast decision-making (68%), 'data leaders' have largely circumvented these issues, with only 8% experiencing similar constraints. This highlights that overcoming legacy data limitations is essential not only for trust but also for the agile deployment and expansion of AI capabilities within an enterprise.

Paving the Way for Future AI Agent Readiness

The increasing pressure to prepare data environments for the age of agentic AI is undeniable. Projections from Gartner suggest that by 2027, AI agents will augment or automate 50% of business decisions, making it imperative for organizations to eliminate data bottlenecks. Without proactive measures to remove these constraints, the promise of enhanced speed and efficiency offered by agentic AI will remain largely unfulfilled. Looking ahead, a staggering 100% of all respondents in the recent executive survey plan to integrate agentic AI within the next two years, with 69% anticipating its widespread adoption across their operations. This universal commitment to agentic AI underscores the urgent need for strategic investment in data infrastructure that can support this technological shift.

To effectively prepare for this future, organizations are prioritizing key initiatives. Improving AI agents' access to both structured and unstructured data stands out as the most crucial step for enabling scalability. Concurrently, enhancing data and AI governance by embedding relevant business context is also high on the agenda, ensuring that AI decisions are not only data-driven but also align with strategic objectives. 'Data leaders' are further distinguishing themselves by focusing heavily on automating data management processes, a move that streamlines operations and ensures data quality and availability for their AI agents. These strategic priorities collectively form a comprehensive approach to building an 'agent-ready' data estate, crucial for capitalizing on the transformative power of agentic AI and securing a competitive edge in the evolving business landscape.

A New Direction: Amazon's Evolving Gaming Strategy

The Winding Down of MMO Operations

Amazon's venture into the MMO gaming market is drawing to a close, with the company initiating a phased withdrawal from managing prominent titles. Following an announcement last year regarding a substantial reduction in its first-party AAA game development, particularly within the MMO genre, Amazon is now formalizing the transition of its existing MMO portfolio. This strategic move impacts two key games, Throne and Liberty and Lost Ark, both of which will soon operate under new management in Western markets.

Transitioning Ownership: Throne and Liberty and Lost Ark

The operational oversight for Throne and Liberty is slated to be transferred to FirstSpark Games and NC during the fourth quarter of 2026. Shortly thereafter, in early 2027, Smilegate will assume responsibility for the live operations of Lost Ark. These transitions signify Amazon's deliberate divestment from the direct management and support of these large-scale online games, allowing external studios to take the helm and continue their development and community engagement.

The End of New World: Aeternum

Parallel to these transfers, another of Amazon's MMO titles, New World: Aeternum, ceased receiving new content updates in October. The game is no longer available for purchase and is scheduled to be permanently taken offline by January 31st, 2027. This complete discontinuation underscores the extent of Amazon's re-evaluation of its MMO commitments, effectively closing a chapter on its direct involvement in certain segments of the genre.

Amazon's Future in Gaming: Cloud, Party, and Franchises

Moving forward, Amazon's gaming endeavors will concentrate on different strategic pillars. The company is intensifying its investment in its Luna cloud gaming service, aiming to provide a seamless streaming experience for a wide array of games. Additionally, Amazon is focusing on developing interactive party games that can be enjoyed on televisions with mobile devices, incorporating innovative elements such as AI-powered characters voiced by celebrities like Snoop Dogg and Arnold Schwarzenegger. A significant part of this refined strategy also involves leveraging established intellectual properties and franchises, creating games that tie into Amazon's broader entertainment ecosystem, including projects related to James Bond and Tomb Raider.