Thoughtworks has released the 33rd edition of its Technology Radar, a biannual report that draws on its global consulting experiences to identify emerging technology trends. The latest volume highlights the rapid evolution of AI assistance and how new systems, protocols, and practices are shaping enterprise software development.
The report observes a clear shift from early experimentation towards maturity, as organisations move from exploring generative AI to implementing it at scale. This edition spotlights growing consolidation around what Thoughtworks calls ‘context engineering’, the adoption of new protocols such as Model Context Protocol (MCP), and the rise of agentic systems. The previous edition, released in April, focused on retrieval-augmented generation and prompt engineering—showing just how quickly the AI landscape is changing.
From AI hype to large-scale adoption
According to Thoughtworks, enterprises are increasingly embedding AI into their operational frameworks and engineering processes. This change reflects a move from proof-of-concept experiments to industrial-scale adoption across sectors.
Sarah Taraporewalla, Chief Technology Officer for APAC at Thoughtworks, said the trend is especially visible in Singapore. “This volume of Thoughtworks’ Technology Radar perfectly captures the shift we are seeing with our clients in Singapore: the decisive move from AI hype to industrial-scale adoption,” she said.
She added that the maturity of AI adoption is most evident in banking and the public sector, where organisations are placing strong emphasis on risk management and governance. “We see this maturity in key sectors like banking and the public sector, where rigorous risk management frameworks and proactive controls are prioritising governance. This same discipline extends to infrastructure; in line with the country’s National AI Strategy, the focus is squarely on optimising high-cost GPU resources. The Technology Radar confirms that rigorous engineering and proactive governance are not barriers to innovation. In a high-trust market like this, they are the essential foundation for building secure, efficient and value-driven AI,” Taraporewalla said.
Key trends in Technology Radar Volume 33
Among the emerging themes, the report highlights the increasing role of agents supported by the Model Context Protocol (MCP). MCP has become a popular integration protocol that enables agents to operate semi-autonomously and efficiently, laying the groundwork for agent-assisted workflows in software engineering.
The Radar also notes that infrastructure orchestration has become vital to managing the surge in AI workloads. With teams needing to coordinate large fleets of GPUs for both training and inference, GPU-aware orchestration is now essential for competitiveness.
Another focus is the evolution of AI-assisted coding. The report finds that AI is embedding itself across the software development lifecycle, helping teams manage context and share knowledge. However, alongside these advances, Thoughtworks warns of a rise in AI-related antipatterns such as AI-accelerated shadow IT and overreliance on machine-generated code. The consultancy stresses the need for continued human oversight and critical judgment to avoid long-term issues.
A call for disciplined AI innovation
The overarching message of the latest Radar is that while AI can drive major efficiency and innovation gains, success requires rigour and careful management. Thoughtworks cautions that the infrastructure, governance, and cultural dimensions of AI adoption must not be overlooked.
The firm continues to advocate a balanced approach that combines technical excellence with ethical responsibility. As AI becomes an integral part of enterprise workflows, organisations must ensure that human decision-making remains central to how systems are designed, deployed, and maintained.
