Let's build together | PromptQL Maxxathon

PromptQL Logo
PromptQL Logo
Sign in
Basics

Approve and collaborate

PromptQL surfaces what it picked up from how your team works. You approve or refine — the semantic layer sharpens with every loop.

What it is

As you and the team use PromptQL, it notices patterns — your definition of top SKUs, the team that owns supply follow-ups, the way you compare weeks. Instead of guessing forever, PromptQL surfaces those patterns as proposed wiki entries and asks you to confirm or edit.

What you approve goes into the semantic layer — a living wiki of how your business actually thinks. Every future thread, artifact, schedule, alert, and action starts smarter because of it. The other four hand-offs in Basics automate the loop; this one is how the loop gets better.

A PromptQL thread where Alex tags @sarah to lock down what 'top SKUs' means; Sarah replies @PromptQL with the team's actual rule; PromptQL thanks her and drafts a structured wiki entry in a 'PromptQL wants to learn' card with Add to wiki and Edit buttons.

How to do it

  1. Keep working. Learnings come from how the team naturally talks to PromptQL — definitions someone drops in passing, corrections mid-thread, terminology used repeatedly across rooms. Pull a teammate in when PromptQL needs to learn something from the actual owner; @-tag them and let them clarify directly in the thread.
  2. When PromptQL spots a pattern worth keeping, it posts a PromptQL wants to learn card in the thread with a draft wiki entry — usually two or three bullets that capture the gist. Anyone in the thread can see what's being proposed.
  3. Approve, edit, or reject. Add to wiki commits the entry as-is; Edit opens the bullets so you can tighten the wording or add a nuance PromptQL missed. Reject and the proposal disappears.
  4. What's approved feeds the semantic layer immediately. The next time anyone asks about top SKUs — in any room, any thread — PromptQL uses your definition without anyone re-teaching it.

What's next

That's the loop, end to end. You asked, PromptQL answered. You saved the answer as an artifact, scheduled it, alerted on it, wired actions to it — and along the way, taught PromptQL what your team actually means by the words you use.

The Basics section ends here. The next layer up is Use cases — full workflows that compose these primitives into something a real role does on a Monday morning.