This Tech Executive Built an AI Assistant That Tells Her to Skip Her Own Meetings
What You'll Find In This Article
- •Understand how to approach building a personal AI assistant without needing programming skills
- •Recognize the value of 'disposable' AI tools built for experimentation rather than perfection
- •Learn a proven strategy for driving AI adoption across a team or organization
- •Identify specific use cases where AI can provide value in managing executive-level workloads
Ever feel like you're drowning in meetings, emails, and calendar chaos? Rachel Wolan, the Chief Product Officer at Webflow, felt the same way—so she built herself a custom AI assistant that acts like a brutally honest chief of staff. Unlike a human colleague who might sugarcoat things, her AI will straight-up tell her when she's wasting time in meetings she should skip or delegate.
The fascinating part isn't just what the tool does—it's how she built it. Wolan treats her AI assistant like a disposable document, not a polished product. She cobbled it together using simple text files and a few software connections, proving you don't need to be a programmer or have fancy technology to create something genuinely useful. Her philosophy? Build fast, learn by doing, and don't worry about making it perfect.
Now she's spreading this mindset across Webflow, running 'builder days' where employees create their own AI tools and requiring people to bring working prototypes to meetings with her. It's a masterclass in how leaders can move their organizations from AI-curious to AI-capable.
The Problem
If you're an executive—or anyone with a packed schedule—you know the drill: your calendar looks like a game of Tetris, your inbox is a black hole, and you spend more time preparing for meetings than actually doing meaningful work. The cruel irony? The more senior you get, the worse it becomes.
Rachel Wolan, who leads product development at Webflow (a popular website-building platform), was stuck in this exact trap. She needed help managing the chaos, but hiring another human assistant comes with its own overhead—and humans tend to be polite. Sometimes too polite to tell you that you're wasting your own time.
The Solution Explained
Wolan built herself a custom AI assistant that functions like a chief of staff—but one with zero filter. This digital helper handles the grunt work (prepping for meetings, sorting emails, organizing her calendar) while also serving up what she calls 'brutal truth' feedback about how she's spending her time.
The AI might tell her: 'You don't need to be in this meeting—delegate it.' Or it might draft a message for her to send, reducing the friction of routine communication. It's like having an assistant who's read every productivity book ever written and isn't afraid to call you out.
How It Actually Works
Here's where it gets interesting for non-technical folks: Wolan didn't hire a team of engineers or buy expensive software. She built this herself using surprisingly simple ingredients:
- Markdown files as a knowledge base: Think of these as simple text documents that store information about her preferences, priorities, and context. It's like giving the AI a cheat sheet about how she works.
- API connections: These are basically digital bridges that let her AI talk to her calendar, email, and other tools she already uses.
- Multiple AI models stitched together: Instead of relying on one AI, she combines several to handle different tasks—like having specialists instead of one generalist.
The key mindset shift? She treats this tool as disposable—built for an audience of one (herself). There's no pressure to make it perfect or shareable. If it breaks or becomes outdated, she just rebuilds it. It's more like creating a personal spreadsheet than launching a product.
Real Examples
Calendar Triage: Before a busy week, the AI reviews her schedule and flags meetings where her presence isn't essential. It might say, 'This status update meeting has no decisions requiring your input—send your questions in writing instead.'
Meeting Prep: Instead of spending 30 minutes reviewing documents before each meeting, the AI summarizes the key points and suggests questions she should ask.
Email Management: The AI sorts incoming messages by urgency and drafts responses for routine requests, leaving Wolan to handle only the messages that truly need her personal attention.
Time Audit Feedback: Periodically, the AI analyzes how she's been spending her time and delivers unvarnished assessments—like pointing out she's been in too many low-value meetings or hasn't blocked enough focus time.
The Ripple Effect at Webflow
Wolan didn't stop with her own productivity. She's using her experience to push AI adoption across the entire company through a clever two-pronged approach:
- Builder Days: These are company events where employees get time and encouragement to prototype their own AI tools. Participants have called the experience 'eye-opening'—many discovered they could build useful tools without being programmers.
- The Prototype Mandate: Want a meeting with the CPO? Come with a working prototype. This simple rule has dramatically increased the number of people actually experimenting with AI rather than just talking about it.
It's a smart combination of top-down pressure ('I expect you to try this') and bottom-up excitement (prizes, recognition, and the genuine thrill of building something useful).
Identify your biggest time-waster: Is it meeting prep? Email overload? Calendar chaos? Pick ONE problem to solve first.
Create a simple text document listing your priorities, common meeting types, and how you prefer to spend your time. This becomes your AI's cheat sheet.
Choose an AI tool to start with. ChatGPT, Claude, or Google's Gemini all work. Copy your priorities document into a conversation and ask for help with your chosen problem.
Test it on a real task: Paste a meeting agenda and ask the AI to summarize key points and suggest questions. See if it's helpful.
Explicitly ask the AI for honest feedback. Describe how you spent your week and ask it to identify meetings you could have skipped.
Iterate and expand. If it's working, gradually add more context to your cheat sheet and try tackling additional problems.
PROMPT:
"Here's my calendar for next week. Based on my priorities [list them], which meetings should I consider delegating or declining?"