Danny De Vries
Essay · Scale Dashboard · Day 1 · Jun 28, 2026

Feed AI Your Inputs, Not the Internet

by Danny de Vries

Here's the thing nobody tells you when they hand you a list of AI prompts: the answers that actually matter to your business aren't on the internet. They're sitting in your inbox, in a voicemail you half-typed up, in the note you scribbled after a job. The most useful thing AI can do for you this week isn't write a blog post — it's read your mess and hand it back clean.

Most small-business owners I talk to are quietly drowning in retyping. A customer sends a rambling email; somebody reads it, then types the name, the address, the phone number, and what they want into a spreadsheet or a job board. Every email. Every call. It feels like nothing — two minutes here, three minutes there — until you add it up.

And it does add up. One professional-services firm found its people were losing about 45 minutes a day each just hunting for the right information — roughly $180,000 a year — and that 41% of their files were duplicates or out of date (AI READY, Danny de Vries). That's the invisible cost of letting your own data live as a wall of text instead of clean records you can use.

The win isn't a smarter prompt. It's pointing AI at the inputs you already have and getting structured data back.

Why this is the foundation, not a nice-to-have

Think of it like prepping ingredients before you cook. A good kitchen doesn't start cooking the second an order comes in — someone has already washed, chopped, and laid everything out, so the line can move fast and nothing gets dropped. Your business data is the same. If every customer message arrives as a raw, unsorted pile, every downstream step — replying, quoting, scheduling, following up — starts from scratch. Clean the input once, and everything after it gets faster.

That's also why this is Day 1 of the challenge. You can't draft a good reply, build a quote, or send a smart follow-up on top of a mess. Capture comes first. Get the inputs clean and the rest of the dashboard has something solid to stand on.

The catch — and the rule that keeps you safe

There's a failure mode worth naming. The book I wrote on this calls it the readiness blindspot: most teams just automate the existing mess instead of rethinking it. In fact only about 21% of companies actually redesign how work flows when they bring in AI (AI READY, Danny de Vries) — the rest pave the same broken path, faster.

So here's the operator rule. The tool reads; you decide. What comes back is a draft. You check the fields — especially anything that's a price or a promised date — before it goes into your system or in front of a customer. And the only thing the tool ever touches is the text you paste. State that plainly, keep a human on the high-stakes calls, and you get the speed without the risk.

You don't have to learn to code to make this happen. You describe what you want in plain English — "pull the name, phone, address, and what they need out of this email" — and you get working structure back. That's the whole shift. Start with the most annoying retyping job in your week. That's your first clean input.

Install the skill: Inbox → Clean Data →
Paste a messy email, transcript, or note; get one clean, structured record back. Free, runs in your AI assistant, the skill stores nothing — output is a draft you confirm.