Let’s skip the hype.
I’m not here to tell you AI is going to change the world. I’m here to tell you what’s sitting open on my desktop right now and exactly how I use it to get more done before noon than most people get done in a week.
Twelve months ago I was skeptical. Not “I haven’t tried it” skeptical — I mean I had tried it, got a robotic blog post and a weird image of a cat in a business suit, and moved on. That was a mistake.
The shift happened when I stopped treating AI like a party trick and started treating it like a $0/hour operator who never sleeps, never complains, and doesn’t need health insurance. Here’s the exact stack, broken down by task — with the specific prompts I actually use.
1. Email: Stop Agonizing Over Words That Don’t Matter
Tool: Claude
I type fast and blunt. That’s a liability when you’re talking to a partner, a client, or anyone who didn’t grow up communicating through Slack at 11pm.
Claude fixes that in under two minutes. And the reason I use Claude specifically over other tools here is simple: it has the most natural-sounding output of anything I’ve tested. It doesn’t write like a press release. It doesn’t add filler phrases like “I hope this email finds you well.” It just makes your actual writing cleaner.
My workflow: I write the ugly draft — the one that says what I actually mean — and hand it to Claude with one of three specific prompts depending on what I need:
- The Professional Polish: Take a blunt internal memo and tell it: “Rewrite this to be professional, polite, and authoritative for an external partner.” Done in 30 seconds.
- The Tone Softener: When I’ve written something that’s technically accurate but will start a fire, I ask: “Soften this tone. Make it sound collaborative and solution-oriented, not confrontational.” This one has saved more than a few business relationships.
- The Topic Expander: When I’ve got three bullet points and need a full team update, I ask: “Expand on these points, add relevant context, and make the flow engaging.” I get a complete draft to edit down from, which is 10x faster than writing from scratch.
What used to eat 15 minutes of mental bandwidth per email is now a 90-second task. The output still sounds like me. It just doesn’t get me in trouble.
2. Meetings: Stop Taking Notes Like It’s 2003
Tool: Plaud.ai
Here’s the math: if you’re in four hours of meetings a day and spending another hour formatting notes for three different departments, you’re wasting 25% of your working life on administrative nonsense that produces zero revenue.
I use the Plaud Note — a slim hardware recorder that clips magnetically right to the back of your phone — to capture every meeting and call. The app transcribes it afterward with high accuracy, but that’s not the part worth paying for.
The part worth paying for is audience-specific output from a single transcript. The same meeting recording gets reformatted into completely different documents depending on who’s reading it:
- For the C-Suite: An executive template that strips out all the ground-level detail and delivers high-level strategic decisions, financial impacts, and immediate blockers. One tight page. That’s it.
- For the Engineering Team: A technical template that ignores the business case entirely and just outputs hard specs, system requirements, and direct action items. Engineers don’t need the backstory. They need the list.
- For Operations and Managers: A breakdown focused purely on timelines, resource allocation, and who owns what deliverable by when.
One meeting. Three perfectly formatted outputs for three different audiences. Zero time spent manually reformatting. That’s an hour a day back in your pocket, and nothing falls through the cracks because everyone got the version that actually applies to their work.
3. Research and Lead Intel: Multimodal Is the Word You Need to Understand
Tool: Gemini
When people say Gemini is “multimodal,” what they mean in practice is that you can hand it a spreadsheet, a PDF, an image, and a web search query simultaneously and get a coherent, synthesized answer out the other side. Most people are still using it like a slightly smarter Google. That’s leaving serious capability on the table.
Here are the three ways I use it that actually move the needle:
- Lead Fit Analysis at Scale: Before a trade show, I get the attendee list as a PDF or Excel file. I upload it alongside our company profile and ask: “Based on our attached company profile, which 10 companies on this list are the best potential fit for our services, and explain why.” What used to take a researcher half a day to cross-reference takes Gemini about 90 seconds.
- Competitor Recon: I use its web-browsing capability for targeted competitive intelligence: “Find five competitors to [Company X] in [specific industry] and list their primary differentiators and target markets.” Fast, structured, and more comprehensive than anything I’d pull together manually in the same time.
- Visual and Technical Intelligence: This is the one that genuinely surprised me. Upload a technical drawing, a patent illustration, or even a photo of an unfamiliar component and ask: “Identify what this object is, find its common specifications, list typical manufacturers, and show me where this component is primarily used in industry.” You get a full technical briefing in under two minutes. That analysis used to mean hiring a researcher or spending hours in technical forums. Now it’s a prompt.
4. Data Entry and Contracts: The Two Things That Kill Productivity Dead
Tool: ChatGPT
Two specific use cases here, and both pay for themselves the first week.
Structured Data Capture: Lay 20 business cards out on a table after a conference. Take one photo. Upload it to ChatGPT and prompt: “Convert all data in this image into a downloadable CSV file with columns for Name, Company, Title, Email, and Phone.” That’s it. Zero manual data entry. Clean, formatted, ready to import into your CRM.
Same principle applies to physical documents. Got a complex table in a printed technical drawing? Photograph it, upload it, ask for the CSV. What took 45 minutes of manual typing is a two-minute task.
Logic and Math Audits: Before any major proposal goes out, I upload the Excel file and ask: “Double-check all formulas in this sheet for accuracy, and flag any numerical anomalies that look like typos.” Catching a broken formula before a client sees it is worth infinitely more than the 60 seconds it takes to run this check.
Contract First-Pass Review: I’m not a lawyer. I also don’t want to pay $400/hour for one to read a boilerplate NDA I get every other week. I upload the PDF and ask: “Summarize the three biggest risks for my company in this contract. What clauses should I push back on, and what wording should I suggest adding to protect my IP?”
Here’s the important caveat I don’t skip: AI does the first pass, a real attorney reviews anything significant. But the difference is I now send my lawyer a two-page structured summary with specific pointed questions instead of a 20-page document and an open-ended ask. His billable hours dropped significantly. My preparation quality went up. That’s a win on both ends.
5. File Organization: The Task Nobody Does Until It’s a Full-Blown Crisis
Tool: Claude Cowork (Anthropic Desktop App)
This is the newest addition to the stack and the one that gets the most reaction when I describe it. Anthropic recently added the ability for Claude to actually interact with your local file system through their desktop app. Not give you instructions on how to organize it. Actually do it.
Here’s how I use it in practice:
The Digital Janitor: I point Claude at my Downloads folder — which at any given time looks like a landfill of generic screenshots, random PDFs, and CSVs with names like “export_final_v3_FINAL.csv” — and tell it: “Organize this folder. Create subfolders by file type, rename all generic image and document files based on the context of what’s inside them, and delete any exact duplicates.” I walk away. I come back to a clean, logical structure where I can actually find things. It read the files, understood the content, built the folder hierarchy, and moved everything. No clicking required.
Local Knowledge Extraction: This one replaced what I used to pay a junior admin to do. Instead of manually opening a dozen different strategy docs, vendor contracts, or technical specs to hunt down one specific detail, I point Claude at the entire project folder and ask: “Read all documents in this folder and compile a one-page summary of every technical requirement mentioned, noting exactly which document each requirement came from.” I get a sourced, organized brief instead of an afternoon of digging.
Batch Data Processing: If I’ve got a folder full of raw expense reports or data exports, I tell it: “Go through every file in this directory and extract the expense data into a single, clean Excel spreadsheet.” It reads each file, pulls the relevant data, and outputs one clean consolidated document.
That’s the kind of task that used to be a half-day project for someone on the team. Now it’s a prompt and a coffee break.
The Part Where I Tell You to Actually Do Something
You can bookmark this and feel good about reading it, or you can pick one tool, install it today, and point it at whatever is currently eating the most of your time.
The businesses pulling ahead right now aren’t the ones with the best product. They’re the ones operating leaner, moving faster, and not paying human-hour rates for tasks that a well-prompted AI handles in 90 seconds.
Pick your biggest bottleneck. Start there. Everything else compounds from that first win.
What’s the first thing you’re handing off? Drop it in the comments.