Question 1

What does the Grounding Check measure?

Accepted Answer

The Grounding Check evaluates how well an entity in a text can be processed by AI systems.

It measures two things:
- Activation (AKT) → Is the entity unambiguously recognized?
- Citation (ZIT) → Is the entity used as the central reference?

Question 2

What is a "well-activated" entity?

Accepted Answer

An entity is well activated when it is clearly defined and unambiguously identifiable.

Good activation (example): "The Eiffel Tower is a structure in Paris in France."

Why this works:
- clear category ("structure")
- unambiguous classification
- no ambiguity

Poor activation (example): "The Eiffel Tower is a world-famous landmark."

Problem:
- no clear definition
- no classification
- only description instead of explanation

Question 3

What lowers the Activation Score?

Accepted Answer

- no clear definition (no "is a …" statement)
- unclear or generic terms without classification
- multiple possible meanings
- purely descriptive statements without context

Question 4

What is a "well-citable" entity?

Accepted Answer

An entity is well citable when it is clearly recognizable in the text as the main reference.

Good citation (example): "The Eiffel Tower stands in Paris. The Eiffel Tower opened in 1889."

Why this works:
- the same entity is used consistently
- clear focus
- no competition from other main concepts

Poor citation (example): "The Eiffel Tower is among the most famous landmarks alongside the Colosseum and the Statue of Liberty."

Problem:
- multiple equally weighted entities
- focus is split
- no clear main reference

Question 5

What lowers the Citation Score?

Accepted Answer

- multiple equally weighted entities in the same text
- frequent shifts of focus
- comparison or list content without a clear main entity
- content without a recognizable priority

Question 6

Why are Activation and Citation separated?

Accepted Answer

Because AI systems take two steps:

1. Understand → What is this?
2. Select → What gets used?

Many pieces of content fail exactly here:
- well explained, but not focused → high AKT, low ZIT
- strongly focused, but unclear → low AKT

Question 7

How can I quickly tell whether a text works well?

Accepted Answer

Check three things:

1. Is there a clear definition at the beginning?
2. Does the text stay with one main entity?
3. Are unnecessary side entities avoided?

If the answer to any of these is "no", the score drops.

Question 8

What kind of content typically performs well?

Accepted Answer

High scores:
- clearly defined entities
- structured facts
- content with an unambiguous focus

Low scores:
- texts covering many topics simultaneously
- comparison or list content
- content without a clear definition

Question 9

What is the goal when optimizing?

Accepted Answer

Not more text. Not more keywords.

Instead:
- clear definition of an entity
- unambiguous focus on one main reference
- consistent structure without ambiguity

This is the foundation that makes content recognizable and usable for AI systems.

Question 10

How many entities can be strong on the same page at the same time?

Accepted Answer

In practice:
- A page has one dominant main entity
- Other entities are usually subordinate

Why: AI systems prefer clear structures.

When multiple entities appear equally strong:
- the scores get distributed
- no entity is clearly prioritized
- citation drops

Question 11

What happens when multiple entities are recognized?

Accepted Answer

That is normal.

The important distinction:
- one main entity → should have the highest values
- side entities → enrich the context

It becomes problematic when:
- multiple entities have similar scores
- no clear hierarchy is visible

Then the content appears unclear to AI systems.

Question 12

How do I recognize a strong main entity?

Accepted Answer

A strong main entity has:
- the highest Activation Score
- the highest Citation Score
- clear presence throughout the entire text

Typical pattern:
- defined at the beginning
- consistently picked up again
- remains the focus of the content

Question 13

What are "weak" entities in a text?

Accepted Answer

Weak entities are terms that:
- are mentioned but not defined
- appear only once or in passing
- have no clear role in the content

They typically have:
- low Activation Scores
- very low Citation Scores

Question 14

Should I remove weak entities?

Accepted Answer

Not necessarily.

There are two sensible options:

1. reduce, if they disrupt the focus
2. expand, if they are relevant

What matters: every important entity needs a clear role.

Question 15

When is it worth creating a dedicated Grounding Page for an entity?

Accepted Answer

When an entity:
- is relevant to the text
- but has low scores
- is not clearly defined

Then it is a candidate for its own page.

Typical signal:
- mentioned but not explained
- has no stable classification

Question 16

How can I use the analysis concretely for that?

Accepted Answer

Proceed like this:

1. Look at all recognized entities
2. Identify entities with:
- low activation
- low citation
3. Check: → Is this entity important for my topic?

If yes: → create a dedicated, clearly defined page

Example

Text contains:
- "Pizza" (strong)
- "Tomato sauce" (weak)
- "Mozzarella" (weak)

Interpretation:
- "Pizza" is the main entity
- "Tomato sauce" and "Mozzarella" are candidates for dedicated content

If they are strategically important: → create dedicated, clearly structured pages

Question 17

What is the goal when there are multiple entities?

Accepted Answer

Not: → squeeze in as many entities as possible

Instead: → create a clear structure

That means:
- one main entity per page
- clear secondary roles for other terms
- no equally weighted competition

This increases the likelihood that content is correctly understood and used by AI systems.