Thoughtworks Unveils 2025 Technology Radar
On November 5, 2025, Thoughtworks, a premier global technology consultancy, announced the latest findings in
Volume 33 of its biannual Technology Radar. This report highlights the significant shifts and rapid advancements in AI assistance within the enterprise landscape, reflecting the observations drawn from extensive client interactions and industry insights.
In this latest edition, Thoughtworks signals a marked evolution from the previous volume, emphasizing the concept of
context engineering and a growing focus on
agentic systems. Unlike Volume 32, which explored emergent practices like retrieval augmented generation and prompt engineering, this year's report showcases a more defined trajectory in the cloud of AI evolution. The pace at which AI is maturing in software engineering is not only notable but also essential for professionals across the industry.
Key Insights and Trends
Thoughtworks identified several core themes in this latest Radar, shedding light on the innovations and challenges influencing the integration of AI in software development:
1.
The Rise of Agent Awareness: With platforms increasingly adopting
Model Context Protocol (MCP), the notion of agent-aware technology has gained traction. MCP serves as an integration protocol that enables agents to operate efficiently and semi-autonomously, thereby reshaping collaborative workflows for software development teams.
2.
Infrastructure Orchestration for AI: The rapid growth in AI workloads necessitates advanced orchestration techniques particularly for managing large fleets of Graphics Processing Units (GPUs) utilized in both training and inference phases. This need for orchestration drives competition within the tech industry to implement
GPU-aware solutions that can adeptly handle these responsibilities.
3.
Integration of AI in Coding Workflows: AI is becoming a fundamental element throughout the software development chain. Teams are now strategizing deliberate methods to manage context with coding agents and effectively share knowledge across projects, improving overall collaboration.
4.
Emerging AI Antipatterns: With the proliferation of AI tools comes the emergence of new issues. These include challenges like
AI-accelerated shadow IT and complacency regarding the quality of AI-generated code. Organizations must prioritize rigorous human oversight and maintain a critical eye on AI outputs to avoid negative ramifications in their ongoing projects.
Changing Perspectives on AI
Rachel Laycock, Chief Technology Officer at Thoughtworks, remarked, “This volume signifies a pivotal shift in industry perception regarding AI. In earlier discussions, concepts like
vibe coding reigned, where developers relied on intuitive prompts for expected outcomes. That trend has diminished significantly. Today, we are witnessing a dedicated effort to navigate the challenges related to context, infrastructure, and security. This ongoing year has acted as a massive live experiment in AI application, and we are beginning to observe tangible benefits emerge from it.”
The Thoughtworks Technology Radar Volume 33 invites industry leaders, technologists, and developers alike to adapt and align with these evolving paradigms of AI. The report serves as a critical resource for those aiming to enhance their understanding and deployment of AI technologies.
For further details, you can visit
Thoughtworks Radar to explore the interactive version or download the comprehensive PDF report. With insights that encapsulate the macro trends impacting the tech landscape, this edition promises to be instrumental in guiding future strategies related to AI in technology practices.
In conclusion, as AI technology continues to advance, it underscores the importance of both innovation and caution within the industry. The findings of Volume 33 reinforce the narrative that while AI can facilitate substantial improvements in productivity and efficiency, the necessity for thoughtful application and robust governance is essential to capitalize on these advancements effectively.