Focusing on supporting underwriting teams by mitigating repetitive manual tasks and accelerating funding decisions for businesses, Ada, the newly rolled out AI agent, aims to advance automated credit decisioning. Developed in-house by Liberis’ AI development teams, the AI underwriting agent integrates and analyses financial, compliance, and digital footprint data in real-time. Afterwards, Ada automatically surfaces key risk factors, trends, and anomalies to assist human underwriters. Additionally, the solution learns from underwriter feedback and augments decision accuracy over time.
Since 2019, Liberis has been investing in AI, with 85% of its financing decisions already being automated. Yet, manual underwriters are leveraged for more complex cases and in these scenarios, underwriting accounts for approximately 48% of application processing time. By launching Ada, Liberis seeks to directly mitigate these inefficiencies by automating data analysis across several sources, including Open Banking data, Companies House records, credit bureau information, and digital footprint indicators.
The system’s architecture includes five integrated layers:
Data ingestion, where raw data is standardised, validated, and enriched;
Feature engineering, with Ada converting this data into structured signals utilised for risk assessment;
Decision support, with the AI agent leveraging business rules and model outputs to generate risk summaries, suggested actions, and confidence scores;
Human feedback loops, with underwriters reviewing Ada’s recommendations and making the final decision. Interactions are then recorded to improve business logic and model parameters;
Delivery systems where key insights are surfaced through Liberis’ Vector platform.
Every day we send out a free e-mail with the most important headlines of the last 24 hours.
Subscribe now