The AskedWell Manifesto
By Paulo de Vries · 2026-05-20
Most websites publish first and hope. They guess at what readers want, write a thousand thin pages, and pray that Google ranks them. The result: the web is drowning in template-padding, AI-slop, and articles nobody asked for.
AskedWell flips the order. We wait to hear what you want. Then we answer as well as we can.
The mechanism
Every page on AskedWell exists because real attention signals earned it into existence. We measure four kinds of asking:
- Search queries. What humans type into Google, what they click, what they scroll past.
- LLM crawl patterns. Which questions GPTBot, ClaudeBot, PerplexityBot, and Google-Extended pull our content for.
- Reading depth. Time on page. Scroll depth. Return visits. Which paragraphs get re-read.
- Sharing + corroboration. What gets quoted, linked, or cited elsewhere on the open web.
When a question gets asked often enough, by enough different signals, by enough different people and bots — we write the deepest answer we can find, with sources cited at every claim. Then we publish.
That's why each page here says “earned its place.” The asking made the page. The page is the asking, answered well.
What we won't do
We've seen what bad publishing looks like. We refuse:
- Template-padding. If a page can't be written by a real person with real research, it doesn't ship.
- AI-slop. Every page must survive a quality bar before publish. AI helps us draft; humans + sources verify.
- Dark patterns. No popups that block reading. No fake-scarcity. No confirmshaming.
- YMYL drift. We deliberately stay clear of medical, financial, and legal advice. Wrong answers in those fields hurt people.
- Verdict labels. No “BUY this stock” or “take this dose.” We're a reference site, not an advisor.
Why this matters in 2026
AI Overviews now eat 50–90% of search clicks for extractable queries. Humans increasingly read answers inside ChatGPT, Claude, and Perplexity without ever visiting a source. Two failure modes are killing publishers:
- Sites optimized only for humans-via-search die when AI Overviews replace clicks.
- Sites optimized only for AI training have no direct revenue path.
AskedWell is built dual-axis from Day 1. We're trustworthy enough that AI cites us (citations bring readers back). We're structured enough that humans find us via search. The loop reinforces itself: AI cites → human clicks → reader signals → deeper pages → more citations → more readers.
Who we are
AskedWell is one person — Paulo de Vries — running a small experiment in demand-driven publishing. More on me. Built with the help of AcePilot, an AI co-builder that respects the rules we set.
What earns a page
Roughly: 10+ distinct asks (humans + bots combined) for the same question within 30 days, OR strong LLM crawl pattern indicating the topic is being answered without sources to cite. When either threshold trips, the topic enters our writing queue. Sourced answer ships in 7–14 days.
The list is currently small. It compounds.
Every page earned its place.