Something weird happened.
I haven't opened ChatGPT in three months. Not because I stopped using AI; I use it more now than ever. But the way I use it changed completely.
Here's what changed: AI moved out of the browser tab and into my actual work.
For a year, using AI meant the same dance every time. Open browser. Open ChatGPT. Copy-paste what I'm working on. Explain the context. Wait for an answer. Copy it back. Edit it to sound like me. Repeat for the next section.
Useful? Sure. But exhausting.
Every time I needed help, I had to stop working, switch contexts, explain everything from scratch, and hope the AI understood enough of what I was trying to do.
It was like having a consultant on call. Better than nothing, but still friction every single time.
Then tools like Claude Code showed up.
Claude Code is an AI tool that works inside your files and folders, rather than in a separate browser chat window.
It's AI that sits on my computer, works inside the files I'm already using, and understands context I would otherwise have to explain. It acts like a thinking partner who's always there, not waiting in another tab for me to remember to ask.
I'm not exaggerating when I say this completely changed the way we work at 6 Seeds.
What This Actually Looks Like
Let me show you the difference.
Old way (browser-tab ChatGPT): I'm working on a keynote deck. Get stuck on how to frame a section. Stop working. Open ChatGPT. Copy-paste my slides and notes. Explain what I'm trying to do. Get back something generic. Edit it for half an hour to sound more like me. Copy it back into the deck. Keep working until I get stuck again. Repeat.
New way (AI in my workflow): I'm working on the same deck. Get stuck. Ask: "This transition feels weak, find me data to help me strengthen it." AI tools like Claude Code already sees the file, knows my style from past work, understands the audience context. Suggests three options that fit. I pick one, keep working.
No tab switching, no copy-pasting, no explaining from scratch.
It's the difference between calling a consultant when you're stuck and having a research partner sitting next to you the whole time.
Here's What Changed for Me
Keynote prep used to take me from two days to a week or even two. Now it takes one day.
AI doesn't write my slides and will never do so. The story will always be mine. But it handles the research, finds the data points I need, drafts the structure so I have something to react to. I spend my time on narrative and delivery instead of hunting for statistics.
Client briefs used to mean digging through five PDFs, three research reports, and a mess of notes scattered across my desktop. Now that gets synthesized in minutes. I still make the strategic calls, but the AI does the extraction and initial structuring. I spend my time on judgment instead of archaeology.
Writing moves faster because I'm not stopping every three minutes to fact-check or find that stat I half-remember. AI pulls it, verifies it, formats it. I stay in flow.
The pattern is the same across everything: AI handles the busywork tax and I handle the thinking.
The quality of my thinking improved because I'm spending more time on strategy and less time wrestling with logistics that break my flow.
Even if you never use Claude Code itself, this shift toward AI working inside your workflow, not in a separate tab, is the signal worth paying attention to.
This Matters if You're in Food and Ag
Saskia and I work with commodity boards and agribusinesses. Every team we talk to is drowning in the same operational friction.
-
•
Research is scattered across decks and PDFs.
-
•
Policy updates arriving constantly but never getting synthesized.
-
•
Retail data that needs formatting before anyone can make sense of it.
-
•
Internal reports that take three days to compile and are outdated by the time they're done.
The bottleneck is turning information into decisions fast enough to matter.
AI eliminates the gap between "I know what needs to happen" and "I have time to actually do it."
When AI works inside your files instead of waiting in a browser tab, it can pull insights from multiple research reports and structure them for decision-making. It can monitor trade and policy developments, flag what's relevant to you specifically, draft the internal update. It can turn messy meeting notes into clear action items without you spending 20 minutes cleaning them up.
We're all drowning in content already, so this isn't about producing more words or AI slop.
This is about executing faster, making decisions quicker and shipping work sooner. Focusing our time on judgment instead of formatting.
The Real Advantage Isn't Content
AI already collapsed the cost of producing content.
When anyone can generate decent copy in seconds, quantity stops being a strategy. "Posting more" is just panic at this point.
The advantage goes to teams that can move faster from insight to decision to action.
And workflow AI, the kind that sits inside your files instead of waiting in a browser tab, changes execution velocity completely.
-
•
I've closed deals faster because proposals get drafted in hours instead of days.
-
•
I've delivered better keynotes because I'm not spending half my prep time hunting down statistics.
-
•
I've made sharper strategic calls because I have more time to think and less time digging through research.
All of this happened because AI removed the friction between knowing what needs to happen and actually getting it done.
What to Actually Pay Attention To
If you're trying to figure out how AI fits into your work, here's what I've learned:
-
•
Ignore tools that just produce more content. You don't need more words. You need better decisions.
-
•
Pay attention to tools that work inside your actual workflow—your files, your documents, the stuff you're already doing every day.
-
•
Ask yourself: where am I losing time to busywork that doesn't actually require my expertise? That's where AI creates leverage.
-
•
Start small. Pick one recurring task that's high-frequency and low-joy. Monthly reporting. Research synthesis. Competitive scans. Use AI to draft, structure, and consolidate it. Keep judgment human.
You don't need to overhaul how you work to benefit from this shift. You just need to remove friction in one place where it slows you down today.
The point is about removing friction so the work that actually matters - strategy, relationships, judgment - gets more of your time.
Getting Started With Claude Code (The Basics)
-
•
Get access and install it
Claude Code is installed on your computer and connected to your Anthropic account. It doesn't run in a browser like ChatGPT.
-
•
Run it where your work lives
You choose which folder or project Claude Code can see, such as a research folder, a report, or a deck. It only works inside the files you point it to.
-
•
Start with one repeat task
Use it first for something simple and recurring like summarizing research, structuring a report, or organizing briefing notes. Let AI handle the setup and synthesis. Keep decisions and judgment human.
This Shift Is Already Happening
AI went from novelty to tool in about 18 months.
Now it's moving from tool to workflow partner.
Most people are still using it the old way. Opening a chat window, asking questions, copying answers back into their work.
But the teams that figure out how to embed AI into their actual workflow, not as a separate conversation but as continuous support, are going to move faster, decide better, and execute cleaner than everyone still context-switching.