Anthropic, the US-based AI startup, finds that 79 percent of conversations on Claude Code, its specialist coding “agent,” have been identified as “automation.” In automation, the AI model performs all the tasks completely, rather than “augmentation,” where AI collaborates with and enhances human capabilities. On the other hand, only 49 percent of the conversations on Claude.ai have been classified as automation tasks. 

The company analyzed around 500,000 coding-related interactions across Claude.ai and Claude Code to find out how people are taking the AI revolution in their work spheres. The increasing reliance on AI suggests that the industry might see greater adoption and involvement of AI as the agentic era unfolds. In tasks, user interface and user experience have been among the top coding uses. 

Scope of AI Involvement 

Out of the 79% of automation tasks served to Claude code, 35.8% of interactions were “Feedback Loop,” where the model completes tasks autonomously but with the help of human validation. On the other hand, it made only 21.3 percent on Claude.ai. “Directive” conversations, where Claude completed a task with minimal user interaction, made 43.8 percent on Claude Code and 27.5 percent on Claude.ai. 

Learning and other augmentation patterns appeared far less often on Claude Code than on Claude.ai

To read about Google's latest addition in Music AI Sandox, click here! 

Mostly Used for? 

The analysis shows that developers often rely on Claude for creating user interfaces and interactive components for web and mobile applications. Though no single programming language prevailed, JavaScript and TypeScript together comprised 31% of all queries, while HTML and CSS contributed another 28%, highlighting a strong focus on front-end development.

Emergence of an Agentic Approach 

The aforementioned data suggest higher usage of the agentic model rather than the general AI model. As more agentic AI tools are released, we may see shifts in how AI is integrated into the workplace, particularly in coding, where this could lead to greater task automation. As AI takes on more component creation and styling tasks, developers may shift their focus to higher-level design and user experience work. 

To read more about Perplexity's latest entry into the WhatsApp landscape, click here! 

AI is fundamentally transforming how developers work, especially when using specialized agentic systems like Claude Code. The analysis suggests these tools are particularly effective for user-facing app development and may offer a distinct advantage to startups over more established enterprises.