For three months, I became a human stopwatch. Every single minute of my day, from the moment I woke up to the second I fell asleep, was tracked in a ridiculously detailed spreadsheet. I’m talking about everything: 7:32 AM - 7:38 AM: Brush teeth and stare into the void. 11:04 AM - 11:05 AM: Respond to a one-word text. It was insane. And the results were, frankly, depressing.
I’ve built and sold two companies. I’ve invested in over 200 startups, including some of the biggest names in AI like Anthropic and OpenAI. You’d think I’d have my time management figured out. Nope. The data showed a horrifying truth: I was losing huge chunks of my day to what I can only describe as “digital sludge.” The worst offender? Meetings. Not just the meetings themselves, but the black hole of time that surrounded them: the prep, the follow-up, the endless email chains trying to remember who agreed to what.
It felt like I was constantly wading through mud. I was busy, sure, but was I productive? The spreadsheet screamed “no.” I was spending nearly a full workday every week just on meeting administration. That’s when I got obsessed. I knew AI had to be the answer. The market was already flooded with tools that promised to be my personal assistant, my scribe, my second brain. So I did what any self-respecting, data-obsessed entrepreneur would do: I decided to test them all.
I spent the next eight months and, yes, over $10,000 of my own money, putting the top AI meeting and productivity tools through the wringer. This isn’t another fluffy listicle. This is my brutally honest, data-backed review of what actually works, what’s a complete waste of money, and the one specific trick that genuinely gave me back 3 hours a week.
The $10,000 Graveyard of Bad AI
My credit card statement for those eight months was a bloodbath. I went after every shiny new tool that landed on Product Hunt with a slick landing page and a lofty promise. You have the AI scribes that join your Zoom calls and spit out a transcript. You have the schedulers that promise to eliminate back-and-forth emails forever. Then you have the all-in-one platforms that want to be your entire operating system for work.
I’m not going to name and shame every single one, but let me tell you, most of them are garbage. The transcripts were often a mess of misattributed speakers and hilariously wrong words. I once had a tool transcribe “Scale AI” as “Whale Sail.” Not helpful when you’re trying to remember the details of a due diligence call. I honestly had no idea what I was doing at first, just throwing money at the problem.
Many of the tools created more work than they saved. They required so much configuration, so much manual correction, that I was spending more time managing the AI than I was on the actual work the AI was supposed to be doing. It was a classic case of a solution in search of a problem. One platform, which cost a hefty $200 a month, promised to create a searchable knowledge base of all my meetings. Great idea, terrible execution. The search was awful, and I could never find what I was looking for. It was like a digital attic full of junk I’d never look at again. After two months, I pulled the plug. That was a quick $400 down the drain.
This is the stuff nobody talks about. Everyone wants to sell you on the dream of AI-powered productivity, but they don’t show you the messy reality of implementation. They don’t tell you that most of these tools are just thin wrappers around a large language model, with very little unique technology or thoughtful design. If you're interested in how I evaluate companies, you can read about my angel investing framework which I apply even when testing products.
The Trick Isn't a Tool, It's a System
After months of frustration and a lighter wallet, the answer finally hit me. It wasn't in a meeting, but on a long Saturday morning run with no podcasts, no music, just the sound of my feet hitting the pavement. My mistake was looking for a single, magical tool to solve all my problems. The real secret, the trick that nobody was talking about, wasn't a product. It was a process. It was about chaining together a few, specialized AI tools to create an automated workflow that handled the entire lifecycle of a meeting.
Here’s the thing: most all-in-one platforms are a mile wide and an inch deep. They do ten things, but none of them particularly well. The breakthrough for me was realizing I could get far better results by picking the best-in-class tool for each specific job and making them talk to each other. This is the system that saved me 3 hours a week. It’s not sexy, it doesn’t have a fancy name, but it works.
My Automated Meeting Workflow
Step 1: The Smart Scheduler
It starts with scheduling. I use an AI-powered scheduler that hooks into my calendar. When someone emails me asking for a meeting, I just CC the AI. It handles all the back-and-forth and finds a time that works for everyone. This alone saved me from countless pointless emails. There are a few good ones on the market, but the key is finding one that reliably understands natural language requests.
Step 2: The Silent Scribe
Once the meeting is booked, the scheduler automatically adds a Zoom link. And, crucially, it also invites my AI scribe to the meeting. This tool’s only job is to join the call, record the audio, and produce the most accurate transcript possible. I’ve found that the tools that focus solely on transcription, without all the other bells and whistles, tend to have the highest accuracy. They’re not perfect, but they’re good enough for the next step.
Step 3: The AI Engine (This is the Real Magic)
This is where the trick really happens. The AI scribe automatically saves the transcript to a designated folder. I have a simple automation set up that watches this folder. As soon as a new transcript file appears, it triggers a script that feeds the entire transcript into a powerful large language model. As an investor in both Anthropic and OpenAI, I’ve had the privilege of playing with their most advanced models, and they are astonishingly good at this.
I don’t just ask it to “summarize the meeting.” That’s a rookie mistake. I use a very specific, multi-part prompt that I’ve refined over months. It looks something like this:
*You are an executive assistant for a busy venture capitalist. Your job is to process the following meeting transcript. Your output must be a draft email to the other participants. Do not be overly formal. Be concise. The email must have the following sections:
- Key Decisions: A bulleted list of any concrete decisions that were made.
- Action Items: A table with three columns: 'Action Item', 'Owner', and 'Due Date'. Extract every single task that was mentioned.
- Concise Summary: A two-to-three sentence summary of the main discussion points. Do not write more than three sentences.*
Step 4: The Draft in My Inbox
The output from this prompt isn’t just a summary; it’s a perfectly formatted draft email sitting in my drafts folder, ready to go. The subject line is pre-filled with the meeting title. All I have to do is read it over, make any minor tweaks, and hit send. The entire process, from the end of the meeting to a follow-up email being drafted, takes about 90 seconds and is completely automatic.
This system took the biggest source of my “digital sludge” and completely automated it. The cognitive load of trying to remember action items, of drafting follow-up emails, of organizing notes—it all just vanished. This is the real secret to founder productivity; it's not about working harder, it's about building systems that work for you.
Why This Works (And Why You Should Steal It)
This system is effective because it plays to the strengths of AI. It uses specialized tools for specific tasks and then uses a powerful language model as the “brain” to connect them and produce a useful output. It’s not about replacing human judgment; it’s about augmenting it. I still review the email before it goes out. I still own the relationships. But I’ve outsourced the tedious, soul-crushing administrative work to a machine.
Look, I get it. Setting this up takes a bit of upfront effort. It’s not a one-click solution. But the payoff is enormous. Three hours a week might not sound like a lot, but that’s 150 hours a year. That’s almost four full work weeks. That’s more time to spend with my family, more time to think deeply about new investments, and more time to actually do the work that matters.
Don’t just buy the hype. Don’t throw your money away on a dozen different AI tools that promise the world. Instead, take a step back. Look at your own day. Find the “digital sludge,” the repetitive, low-value tasks that are eating up your time. Then, get creative. Find the best-in-class tools for each part of the problem and stitch them together. That’s the real AI trick. It’s not about finding the perfect tool; it’s about building your own perfect system.
Now, if you’ll excuse me, my AI just drafted a follow-up email for me. I should probably go and press send.
Frequently Asked Questions
What's the most common pushback you get on this?
People often push back by citing exceptions or edge cases. And they're usually right that exceptions exist. But building a strategy around exceptions rather than patterns is a losing game for most founders.
Do all experts agree with this view?
No, and that's fine. The best ideas in business are often contrarian. I share my perspective based on my experience and data, but I encourage you to seek out opposing viewpoints and form your own conclusions.
What experience informs this perspective?
This perspective comes from over a decade of building companies in Silicon Valley, two successful exits (RemoteTeam to Gusto, MovieLaLa to Gfycat), and investing in 200+ startups including Anthropic, OpenAI, and Scale AI. I write about what I've lived.
How can I apply this thinking to my own situation?
Start by identifying the core principle behind the opinion, not the specific example. Then ask yourself: does this principle apply to my context? If yes, test it in a small, low-risk way before going all in.