Skip to content
NextLine

How it works

A loop that never stops.

NL008 — the 8 is our loop. Observe, interpret, act, re-measure, and go round again. Every turn leaves evidence behind.

The loop

01Observe02Interpret03Act04Re-measure
  1. 01

    Observe

    Ask the engines the questions your customers ask. Store every answer as evidence.

  2. 02

    Interpret

    Find mentions, recommendations and citations — and what is missing or wrong.

  3. 03

    Act

    Ship truthful fixes: clearer facts, better sources, approved outreach.

  4. 04

    Re-measure

    Ask the same questions again. Keep only what the evidence supports.

Evidence chain

Every number can be traced.

Any mention in any report can be followed back to the stored answer, the adapter and parser versions, and the exact rules that fired.

  1. 01

    Question

    A real question, in a set, with engine, region and language fixed.

  2. 02

    Run

    The engine is asked through a versioned adapter.

  3. 03

    Snapshot

    Raw answer, citations and timestamp are stored.

  4. 04

    Parse

    A versioned parser records mentions and rule hits.

  5. 05

    Review

    People can override — with a reason, logged.

  6. 06

    Report

    Figures link back to every step above.

Engine compliance

We follow each engine’s rules.

Every engine is placed in a tier that decides what may be automated. If an engine does not allow it, we do not do it.

AAutomated runs allowed.

Level A · Official API

Where an engine offers an official API, we use it — within its terms.

BRate-limited and allow-listed.

Level B · Limited automation

Where automation is permitted with limits or authorisation, it runs slowly and only on approved questions.

CNo automation. Manual capture only.

Level C · No API or sensitive terms

Where terms do not allow it, nothing is automated. Answers are captured by people and recorded as such.

Content principles

What we publish — and what we never will.

Every piece of outreach and published content is checked against these lines before it is approved.

  • Only true, verifiable content — every claim marked with its source.
  • Advertising is always labelled as advertising.
  • No mass-produced articles, no ghost-written reviews or testimonials.
  • No instructions hidden in pages for AI systems to read.
  • Extra review for clients in medical and health fields.
  • No promise of guaranteed placement — ever.

Ready when you are.

Tell us about your brand and the questions people already ask AI in your category.