AI Agents

Context

Built for banking as a Stand-alone Quick Win and owned by Compliance, this agent automates adverse news checks for entities as part of KYC and periodic reviews. It’s aimed at Tier 1–2 banks that today rely on manual Google/Lexis searches whenever a new or existing counterparty needs a basic adverse media check.

What it does

The agent takes the entity (and, where applicable, related group entities or key principals) and runs targeted adverse media searches across open-source and licensed news/content feeds. It normalises names and key identifiers, applies focused search patterns, and returns a concise view of potential adverse mentions instead of pages of raw results. For each entity, Compliance sees whether adverse news was found, what the main themes are, and a short list of supporting articles to review, so they can make a documented judgement quickly.

Core AI functions

Using OSINT and NLP, the agent gathers news and public content about the entity, filters out obvious noise, and clusters what remains into simple categories (e.g., legal issues, regulatory actions, financial distress). It focuses on turning unstructured search hits into a compact, readable signal that Compliance can consume inside existing KYC workflows, rather than as a separate research exercise.

Problem solved

Today, adverse media checks often mean manual Google/Lexis triage: analysts run their own searches, skim headlines, paste links into notes, and each person applies slightly different search terms and thresholds. That is slow, inconsistent, and hard to audit. The agent standardises that step by running structured, repeatable searches and presenting findings in a uniform format, so “we looked for adverse news and here’s what we found” is handled the same way across teams and cases.

Business impact

The primary impact is faster KYC and review cycles. Compliance can cover more entities with consistent adverse media checks, spend less time on repetitive searching, and focus their effort on interpreting a smaller set of curated results. For the bank, that means quicker onboarding and periodic reviews with a clearer record that adverse media screening was performed as part of the risk-based approach.

Integration and adjacent use cases

Integration complexity is Low–Med: the agent needs access to entity data from the KYC/onboarding or review case, connectivity to your chosen news/content sources, and a way to store its adverse media summary back into the same case.

Common combinations in this Stand-alone Quick Win stack:

• Credit Score Improvement Assistant to give customers clear actions to become eligible once risk and reputation checks are clean;

• Alternative Risk Assessor to supplement bureau scores for thin-file customers using alternative data;

• Customer Profiling & Product Recommender to propose relevant next products once a customer passes basic risk and reputation checks;

• Customer 360 Summary Generator to surface the adverse-media signal as part of the RM’s at-a-glance view;

• Inbound Email Agent (RM Copilot) to pick up client questions about onboarding or declines and fold in the latest risk context when drafting replies; and

• Legal Contract Drafting Assistant to generate clean contracts and letters once a customer has cleared the KYC process and is ready for new products.