AI biweekly recap #3 – top AI use cases in 2025

Welcome to the third issue of The AI Biweekly Recap by Fellow!

In this edition, we bring fresh insights from Anthropic’s Economic Index, invite you to an exclusive live event on how to get started with AI Agents in 2025, and share the latest news on AI in the workplace. Let’s dive right in!

Millions of conversations confirm: AI powers programming and penmanship 🤖✍️

Anthropic researchers analyzed millions of user interactions with their Claude.ai platform to understand how AI is actually being used for different work tasks.

Here are three key findings:

1. Software engineering and writing-related tasks dominate AI usage, accounting for nearly half of all usage.
   

2. A notable 57% of interactions suggested augmentative uses, where people iteratively refine their work or learn new concepts with Claude’s help, while 43% indicated more automative patterns, in which tasks are delegated to the AI with minimal human involvement.
   

3. Fully integrating AI across all tasks in an occupation is still rare: only about 4% of occupations show usage of AI in at least three-quarters of their tasks.

This research is presented as the first large-scale study to track how AI is taking on tasks across the economy, offering an early glimpse into how advanced language models are already reshaping work.

Source: The Anthropic Economic Index (February 2025)

[Free virtual event] AI Agents in 2025: Preparing Yourself and Your Team 🚀

Want to learn real-life use cases and pitfalls to avoid with AI? You're invited to a live event and Q&A on how to get started with AI Agents.

Why should you attend?

• Existing use-cases and applications

• How to get started with AI Agents

• What to consider when implementing AI Agents

• What to avoid! ⚠️ 

Our panel of experts:

• Rob Williams, Fractional Chief AI Officer

• Alex Lee, Founder of Kazka AI Consulting

• Aydin Mirzaee, CEO at Fellow

When: Wednesday, February 26th at 12pm EST / 9am PST
Where: Online

Register now: https://fellow.app/live/ai-agents

This week in AI 📰

1. Agentic report generation is poised to become a top AI use case among enterprises this year, with Deep Research by OpenAI proving its efficiency in handling complex data synthesis
   

2. Microsoft research suggests AI could save UK public sector workers 23 million hours weekly by automating administrative tasks, freeing up time for frontline services. 
   

3. Leaders from Google, Meta, and Salesforce have noted that AI is now writing a significant portion of new code, predicting a transformation in programming roles by the end of 2025.
   

4. AI governance to ensure responsible, safe, and ethical AI use in the workplace will be a priority for organizations in 2025, with new regulations expected to build on existing data protection laws. (Source: University of Salford)
   

5. The proliferation of AI in the workplace might pave the way for more discussions around the 4-day workweek, as AI increases efficiency, potentially allowing for reduced working hours without loss of productivity. (Source: Newsweek)
   

What’s your take on these new models and trends? Hit reply and let us know which developments excite you the most. We’d love to hear your insights and questions.

Until next time,

Manuela and the Fellow team