AI Biweekly Recap – AI Agents in 2025 (+ This Week in AI)

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

We just wrapped a fascinating panel on of AI agents in the workplace: “digital employees” that don’t just answer questions but make independent decisions about tasks and tools.

Interested in learning more about AI agents, common use cases, and pitfalls to avoid? You can watch a replay of the conversation here. 

3 Key takeaways from the panel 🎤

1. Defining AI agents vs. traditional automation

AI agents differ from standard automation or chatbots by having “agency.” Instead of merely following predefined steps or Q&A interactions, they can autonomously decide which tools or data to use and perform multiple actions on their own.

2. Real-world use cases and workflows

Some great ways to use AI agents include helping with contracts, reaching out to new customers, and answering support questions. AI agents excel in these areas by handling unstructured data in ways traditional “process automation” can’t. However, according to our panellists, employing narrowly focused AIs for each step helps minimize errors and AI hallucinations, ultimately reducing manual workload and streamlining critical processes.

3. Tools shaping the evolving AI ecosystem

Panelists highlighted favorites like Anthropic’s Claude, ChatGPT’s Deep Search, and Grok, noting how AI continually evolves. They also mentioned workflow automation platforms like Zapier, Make, N8n, as well as low-code builders such as Lovable and Bolt. Coding assistants and collaboration tools like Cursor and Replit were noted, along with experimental builders such as Diffy. Finally, Alex described Fellow as an AI meeting agent that takes meeting transcripts and turns them into actionable information, such as identifying action items, updating CRMs, and summarizing key insights.

Watch the entire conversation on our YouTube channel to see our panelists tackle other great questions from the audience!

This week in AI 📰

• Employees expect AI to reshape jobs, but leaders lag: A recent McKinsey report finds that employees are more prepared for AI adoption than leaders realize. While 94% of employees are familiar with AI tools, leaders underestimate how much AI is already integrated into daily work. However, 41% of employees remain apprehensive, signaling a need for support, transparency, and training alongside AI expansion.

• AI adoption among small and midsized businesses (SMBs): According to Verizon’s latest State of Small Business Survey, 39% of SMBs now leveraging AI tools for tasks like automating repetitive processes, personalizing marketing, and analyzing data for smarter decisions.

• Leaders must encourage open dialogue amongst AI concerns: A recent Pew Research Center survey (published February 25, 2025) reveals that many American workers feel more worried than enthusiastic about how AI might shape their jobs. Although some employees recognize the potential for AI to increase productivity, others fear it could make their roles obsolete.

As leaders, it’s vital to acknowledge these mixed sentiments and create opportunities to discuss both the challenges and benefits of AI with your teams. By engaging openly and offering support such as training, resources, and clear communication, you can help address employee anxieties and foster a more confident, forward-looking working environment.

We’d love to hear your perspective: what actions or resources do you find most helpful in starting conversations with your team about AI?

We hope this issue of the AI Biweekly Recap by Fellow was helpful.

Until next time,

Manuela and the Fellow team