Overview
When a Watchlist Screening check returns a Consider result, Legl's AI-assisted review automatically analyses each potential match and groups results to help your team prioritise their review.
Rather than presenting all results in a flat list, the AI assesses each match against your client's known details and categorises it as either likely to relate to your client or likely to be unrelated. This reduces the time your team spends on false positives and helps focus attention on the results that need attention.
How this works
When a Consider result is returned, the AI reviews each potential match against your client's known details. It produces two outputs for each match: a category and a short summary explaining the reasoning behind it.
Categories
Each match is assigned one of two categories:
Likely the same person: the AI has identified meaningful overlap between the match and your client's profile. These results require careful review.
Likely a different person: the AI has assessed that there are significant reasons to believe this match relates to someone other than your client. These results are still visible and can be reviewed, but are lower priority.
What the AI considers
The AI uses the data available within each screening result to reach its assessment. Factors it considers include:
Name, date of birth, and country of origin
The jurisdiction and language of the source material, for example whether an adverse media article relates to an individual based in a different country
Descriptions and addresses in the source data that do or do not align with your client's profile
Whether corroborating identifiers, such as age or location, match your client
The type and risk level of the match: PEP and sanctions results are given greater weight in the assessment
AI summary
For each match, a short AI-generated summary is displayed within the result. This explains the reasoning behind the categorisation and highlights the key data points used in the assessment. The summary is designed to provide context at a glance, similar to the research a reviewer might otherwise carry out manually before making a decision.
ℹ️ Further guidance
What triggers this
The AI-assisted review runs automatically whenever a Watchlist Screening check is performed and returns a Consider result. No action is required to trigger it.
What you will see
In the Watchlist Screening panel, results are grouped into two sections:
Likely the same person
Likely a different person
Both sections are always visible. All matches returned by the screening check are displayed: nothing is hidden or removed from view.
Opening any individual result displays the AI summary at the top, followed by the full match detail including source links and data points.
If your firm has not yet switched to the AI-assisted view, a blue banner will appear at the top of the Watchlist Screening panel. Select Switch to AI review to enable the AI-assisted categories.
Important information
The AI categorisation is a triage aid only. Your firm must still apply its own risk-based judgement and review all results in the Likely the same person category.
Results in the Likely a different person category are still visible and require a review decision. The AI triage does not replace your firm's review obligation.
The AI can only assess information included in the screening result. It does not have access to external databases or your firm's internal client records.
PEP and sanctions matches are weighted more heavily in the AI's assessment due to their higher inherent risk.
The AI summary is informational only and does not constitute legal or compliance advice. Each firm must apply its own risk-based approach and internal escalation process.