What If Every Employee Had Their Own AI Executive Assistant?

What if every person in your org - yes, even your intern - had an executive assistant that outperformed most humans? That’s the world Nick Sonnenberg is building.

As a former high-frequency trader turned operational efficiency expert, Nick now helps companies reinvent how work gets done through AI-powered delegation, dynamic workflows, and personalized automations.

In this episode of This New Way, Nick pulls back the curtain on how his team uses orchestration agents, context-rich data pipelines, and a living prompt database to drive results faster than ever.

→ Subscribe to our YouTube Channel

→ Listen to Nick’s episode

Keep reading to get the step-by-step tutorials and AI tools mentioned in this episode. 👇

Prevent Shadow IT, Protect Your Organization:

Freemium, unvetted tools are quietly recording sensitive meetings, storing transcripts off-platform, and slipping past security reviews. This practical guide for the secure adoption of AI notetakers gives you the following tools:

• A 3-step Shadow IT Prevention Framework

• AI Usage Policy Template

• Email, Slack, and Meeting Script Templates

• Feature checklist to vet secure AI notetakers

Tutorial: Build Your Personal AI Assistant in Claude

Nick Sonnenberg’s team uses Claude to do what human assistants used to do: only faster, more accurately, and for everyone on the team.

Here’s how to build your own AI-powered assistant:

1. Set up Claude with MCP (Model Context Protocol).
This enables Claude to securely connect with your internal tools: email, calendar, CRM, and more.

2. Feed Claude your context.
Integrate data sources like:

• Calendar (for meeting schedules)

• Email history (for past conversations)

• CRM (like HubSpot, for customer info)

• Project tools (like Asana, for tasks)

• Knowledge base (like Notion or Coda)

3. Define your agents and hierarchy.
Create specialized agents for things like:

• Drafting emails

• Prepping meeting briefs

• Managing tasks or scheduling

Next, use an “orchestrator agent” that acts like a manager, deciding which sub-agent should handle each task.

4. Automate triggers.
When a new email comes in, your orchestrator agent pulls context and assigns it to your email-drafting agent, who writes a personalized reply.

You can also:

• Trigger agents from Claude directly (e.g. “Tell me everything about [Person]”)

• Generate morning briefings for your meetings

• Route incoming tasks automatically

5. Review and send.
You stay in control: scan the output, tweak it if needed, and hit send. Over time, the assistant learns your voice.

Pro Tip: All prompts and workflows are stored in a database, so your team can version, test, and improve them over time - just like real product code.

👉 Watch the full demo here

AI Tools Mentioned

• Claude + MCP – Context-aware agent orchestration
  

• N8N – Automation and agent workflow builder
  

• Perplexity – Real-time external research
  

• HubSpot – CRM context layer
  

• Asana – Task and planning agent integration

AI workflows aren’t just cool—they’re how leaders are scaling impact fast.

👉 Subscribe to This New Way: YouTube Channel

 📩 Was this email forwarded to you? Subscribe here

Until next time,

Aydin Mirzaee
CEO at Fellow & Host of This New Way