How I AI: Zapier EA Cortney Hickey's 4 AI Workflows for Meeting Prep, Culture Reinforcement, and Strategy

This is one of the most practical, time-saving, and stress-saving conversations we’ve ever had on the show. We’re talking about the operational backbone of a company—all the manual, repetitive work—and how AI can turn it into something automated and strategic.

I sat down with Cortney Hickey, the Executive Assistant to the CEO at Zapier. As you can imagine, working at an automation company means she’s on the front lines of using AI to optimize her role. But what Cortney is doing goes way beyond simple task management. She’s building systems that not only give her hours back every week but also scale the effectiveness of the entire executive team and reinforce the company’s culture.

We often hear the fear that AI will replace roles, especially administrative ones. Cortney is living proof of the opposite. For her, AI is a tool for immense leverage. It lets her work herself out of the boring, repetitive tasks so she can focus on higher-impact, more interesting strategic work. It’s a mindset shift from if you should use AI to when and how.

We’re going to walk through four incredible workflows that Cortney has built. We’ll cover how she created an AI agent to handle all her weekly meeting prep, designed an automated feedback system to coach leaders on company values, built a custom GPT to stress-test documents before they get to the CEO, and set up a centralized knowledge base that makes company strategy interactive and accessible for everyone.

Workflow 1: The Automated Weekly Meeting Prep Agent

We’ve all been there: it's Friday afternoon, and you're staring at next week’s calendar, trying to figure out who you’re meeting, why you’re meeting, and what you need to do to prepare. Cortney used to block off two hours every Friday for this exact task. Now, she’s automated nearly the entire process with a custom agent built in Zapier Agents.

This workflow is a perfect example of reclaiming your time by automating a high-value, repetitive task. Instead of manually digging through a dozen different systems, this agent acts as a research assistant, running in the background to deliver everything you need, exactly when you need it.

The Step-by-Step Process

Cortney built this agent to mimic the exact steps she would take manually. It’s a great example of the “progress over perfection” mindset—she started with a simple digest and gradually added more capabilities over time.

1. Scheduled Trigger: The agent automatically kicks off at 8:00 AM every Friday.
2. Calendar Scan: It first reads her Google Calendar for the upcoming week and identifies all meetings that require preparation, specifically filtering out routine internal stand-ups or one-on-ones to focus on external or high-stakes meetings.
3. Multi-Source Research: For each external participant, the agent performs a series of research tasks:

* It conducts a web search to find their current role, industry experience, and any other noteworthy public information.

* It connects to Zapier's CRM, HubSpot, to check for any existing relationship, recent sales notes, or deal status.

* It searches her internal comms, including Gmail and Slack, to find any prior conversations or mentions of the person or their company.

1. Dual Outputs: Once the research is complete, the agent delivers the information in two useful ways:

Actionable Tasks:* It creates individual tasks in her to-do list app, Todoist, for each meeting. Each task is populated with the full research brief and scheduled to appear two hours before the meeting begins.

Weekly Digest:* It sends a nicely structured digest to a Slack channel. This message summarizes all key meetings, highlights the intelligence gathered, and even suggests key areas of focus for the week.

I always tell people when they're starting with an agent like this is 'Progress over perfection.' I started this one with just a quick digest... And then over time I was like, oh, here's something else that might be helpful. Build something basic, see how it works, learn, and then make time to improve it.

What’s so clever about this is how you can use natural language to build and refine it. Using the built-in copilot, Cortney can simply type instructions like, "for each participant, also add a LinkedIn hyperlink within the Slack digest," and the agent updates its own logic. This workflow completely eliminates a recurring two-hour task and ensures she walks into every important meeting fully prepared.

Workflow 2: Reinforcing Company Culture with an AI Meeting Coach

It’s one thing to have company values written down, but it’s another to make sure people are actually practicing them day-to-day. Cortney built a fascinating workflow that turns meeting transcripts into an automated coaching tool to do just that.

This system provides consistent, unbiased feedback to meeting participants, helping them align their behavior with Zapier's cultural norms. It takes the awkwardness out of direct feedback by making it a routine, expected part of the process, delivered by a neutral AI bot.

The Step-by-Step Process

This workflow uses meeting transcripts from their AI note-taking tool, Fellow, and has AI analyze them against a predefined cultural framework.

1. Ingest Transcript: The automation triggers after a meeting concludes, pulling in the full transcript and metadata (like participant list and meeting duration).
2. Apply Filters: It first checks for certain parameters to ensure the feedback is relevant. For example, it might ignore meetings under 10 minutes or those with external participants.
3. Provide Deep Context: This is the magic step. The agent is fed a rich prompt containing the pillars of Zapier's culture:

Company Values:* Core principles like "growth mindset."

Meeting Norms:* Specific expectations for how meetings should be run.

Frameworks:* It includes concepts like "The Five Dysfunctions of a Team" and the importance of being both "demanding and supportive."

Impact Behaviors:* A list of what is expected from employees at Zapier.

Then, she loops back to the team with feedback in two steps:

1. Generate Personalized Feedback: For each participant, the AI analyzes their contributions in the transcript against the cultural context provided. It generates direct, constructive feedback.
2. Deliver via Slack: The feedback is sent as a direct message in Slack to each person. The message clearly states it is AI-generated and focuses on 1-2 specific growth opportunities and 1-2 things they can do differently next time.

Initially, the feedback was too soft. Cortney iterated on the instructions, telling the agent to strike a better balance between being demanding and supportive. The result was more direct and actionable coaching, like, "Address misalignment more directly," or, "Challenge the decision-making speed."

This system does more than just make meetings efficient; it systematically checks if the team is living up to its own standards. It’s an incredible way to scale leadership coaching and make feedback a normal, continuous part of work.

Workflow 3: Stress-Testing Proposals with a Custom "Exec Prop" GPT

We’ve all been there: you have a big proposal or strategy document that you need to get in front of the executive team. How do you know if it's ready? How do you make sure you’ve anticipated their questions? Cortney found that she was often the go-to person for this kind of pre-review, which created a bottleneck.

To solve this, she built a custom GPT in ChatGPT that acts as a digital clone of the executive thought process. It allows anyone in the company to stress-test their documents and get high-quality feedback before ever scheduling a meeting.

The Step-by-Step Process

The reason this custom GPT works so well is the curated knowledge base it's built on.

Build the Knowledge Base: Cortney uploaded a wealth of internal context into the GPT's knowledge files. This included:

* Company strategy memos and the revenue roadmap.

* Examples of well-written proposals (or "tee-ups").

* Documents detailing team norms and decision-making processes.

* Even a doc called "Managing Up to Wade" (the CEO), which explains his communication style and preferences.

User Submits a Document: An employee can upload their draft document and ask for feedback with a simple prompt like give feedback on a tee-up doc.

AI Analysis and Feedback: The GPT analyzes the submitted document against its entire knowledge base. It identifies gaps, unclear points, and areas that might conflict with strategic priorities.

Structured Output: The feedback is incredibly structured and actionable. It provides:

* A quick summary of the document's strengths and weaknesses.

* Specific advice on how to strengthen it (e.g., "surface trade-offs more clearly").

* An example rewrite to demonstrate best practices.

* A "bold coaching question" designed to push the author's thinking further.

This tool has been a huge success, with over 278 employees using it to sharpen their strategy docs. The result is that when documents do reach the executive team, they are clearer, more aligned, and lead to faster decisions. In some cases, the document becomes so clear that the meeting can be skipped entirely—the ultimate win!

Workflow 4: Building a Centralized Strategy Companion with NotebookLM

The last workflow Cortney showed us tackles another common organizational challenge: making company strategy accessible and understandable to everyone. Strategy is often contained in dense documents or scattered across All-Hands presentations, making it difficult for employees to connect their work to the big picture.

Using Google NotebookLM, she created a centralized, interactive strategy companion that anyone at Zapier can use to get answers.

The Step-by-Step Process

This workflow transforms a collection of static documents into a dynamic, conversational knowledge base.

1. Aggregate All Sources: Cortney and her team uploaded dozens of key strategic documents into a single Google NotebookLM space. This included the top-level strategy doc, transcripts from All-Hands meetings, and every organization's strategic action plan.
2. Ask Questions in Natural Language: Employees can now go to this one central place and ask questions directly. Cortney showed a great example by prompting it: "As an executive assistant, how can I contribute to Zapier's 2026 strategy?"
3. Get Contextual Answers: The tool instantly synthesizes an answer based on the source materials, pointing to specific documents to back up its claims. It told her she could contribute by championing clarity, focus, speed, and driving internal AI transformation.
4. Explore Multimodal Content: Beyond just text answers, Google NotebookLM can automatically generate other ways to engage with the material, like creating an AI-generated podcast summarizing the strategy or an interactive quiz to test your knowledge.

This is a huge deal for organizational alignment. It empowers every employee to understand how their individual role contributes to the company's broader goals, turning strategy from a top-down mandate into a resource that everyone can own and interact with.

Final Thoughts

What I love most about these four workflows is how they show the EA role shifting from tactical support to a strategic partner. By using AI, Cortney isn't just making her own job easier; she's building scalable systems that make the entire organization smarter, more aligned, and more effective.

Her work really pushes back against the fear that AI will displace jobs. When you embrace these tools, you can automate the mundane to free yourself up for the creative, human, and strategic work that truly makes a difference. If you’re wondering where to start, take a page from Cortney’s book: find a repetitive task you do every week, and spend that time automating it instead. You’ll be amazed at how quickly the gains compound.

Thank You to Our Sponsors

• **WorkOS**—Make your app Enterprise Ready today
• **Brex**—The intelligent finance platform built for founders

Episode Links

• Watch on YouTube
• Listen on Spotify
• Listen on Apple Podcasts