The Challenge

OCBC manages high-volume recruitment across graduate, early-career and professional hiring programs.

Academic qualifications and previous experience provide useful signals when evaluating applicants, but they do not always reveal the qualities that can distinguish high-potential candidates. Motivation, adaptability, learning agility, problem-solving approach and long-term potential can be difficult to assess consistently from a resume alone.

The challenge for OCBC was to create a more scalable and consistent approach to evaluating candidates while ensuring recruiters could identify high-potential talent without significantly increasing manual review effort.

The Solution

OCBC leveraged Savos by impress.ai to augment its candidate evaluation process with AI-powered candidate investigation and talent intelligence.

Savos goes beyond traditional resume-based screening by using AI-led conversations to investigate the candidate characteristics that are relevant to the hiring process.

Through ScaleScreen, candidates can be engaged in structured conversations that explore areas such as career motivation, problem-solving, adaptability, learning orientation and role suitability.

TalentLens then synthesizes the information gathered during the candidate interaction into structured candidate intelligence, allowing recruiters to compare candidates using a broader and more consistent set of evidence.

This gave OCBC a way to scale candidate evaluation without simply scaling manual recruiter review. Recruiters could use the resulting intelligence to prioritize candidates, focus their attention on the most relevant applicants and make more informed shortlisting decisions.

How AWS Powered the Solution

Savos was deployed using AWS infrastructure to support its AI-powered candidate evaluation and talent intelligence workflows.

Amazon Bedrock provides the generative AI foundation used by Savos to support candidate investigation, interpretation of candidate responses and generation of structured candidate intelligence.

The application workloads run using Amazon ECS and Amazon EC2, with Amazon RDS and Amazon ElastiCache supporting application data and caching requirements. Amazon SQS enables asynchronous communication between components of the solution.

The architecture uses Elastic Load Balancing and Auto Scaling to support variable recruitment workloads and maintain application availability as demand changes. Amazon CloudFront, Amazon Route 53 and Amazon S3 support application delivery and associated infrastructure requirements.

Security and operational capabilities are supported through AWS KMS, VPC endpoints, security groups, NAT Gateway, AWS WAF and Amazon CloudWatch. The solution uses Multi-AZ deployment, including Multi-AZ configurations for relevant data services, to provide resilience and availability.

Together, these AWS capabilities provide the scalable infrastructure required to run AI-powered candidate evaluation workflows across high-volume recruitment programs.

The Impact

The implementation delivered measurable improvements across OCBC’s candidate evaluation workflows:

  • 60% reduction in application review effort
  • 70% faster applicant processing times
  • 30% improvement in identification of high-potential candidates
  • 25% increase in recruiter productivity
  • 2.5x faster shortlisting workflows

Beyond the efficiency gains, the solution enabled recruiters to evaluate candidates using richer contextual information rather than relying primarily on academic credentials, previous experience and resume-based signals.