We've extended the AI Include Strictness setting to Full-Text AI Screening also, giving researchers greater control over how conservatively or liberally the AI makes inclusion decisions during full-text review.

Previously available for Title & Abstract Screening, this capability is now also applied during the Full-Text Screening stage, allowing teams to fine-tune AI behaviour based on the objectives and rigor of their review.

What's New?

The AI Include Strictness setting is now extended to Full-Text AI Screening.

Researchers can choose how strict the AI should be when evaluating full-text articles against the review protocol.

This allows teams to balance:

• Maximising sensitivity by identifying as many potentially relevant studies as possible.

• Increasing specificity by reducing false-positive inclusions.

How It Works

The strictness setting adjusts the AI's decision-making threshold when determining whether a study should be included.

Users can configure the AI to be:

Conservative (Higher Strictness)

The AI applies stricter inclusion criteria and recommends inclusion only when the article strongly satisfies the protocol requirements.

This setting is useful when:

• Inclusion criteria are highly specific.

• Teams want to minimise false-positive inclusions.

• Reviews require a high level of precision.

Balanced (Medium Strictness)

The AI applies a balanced decision-making approach, providing a good trade-off between Recall and Precision.

This setting is suitable for most systematic reviews and evidence synthesis projects.

Liberal (Lower Strictness)

The AI is more inclusive and recommends inclusion whenever sufficient supporting evidence is identified.

This helps maximise Recall by ensuring potentially relevant studies are less likely to be missed.

This setting is particularly useful for:

• Early-stage evidence reviews.

• Targeted Literature Reviews (TLRs).

• Broad scoping reviews.

• Reviews where missing relevant evidence carries a greater risk than reviewing additional articles.

Why This Matters

Different review projects require different screening strategies.

By extending AI Include Strictness to Full-Text Screening, researchers can now configure AI behaviour independently for both screening stages.

For example:

• Use a more liberal approach during Title & Abstract Screening to maximise Recall.

• Apply a stricter approach during Full-Text Screening to improve Precision before final study selection.

This provides greater flexibility while maintaining a transparent and reproducible review process.