"Operationalising GenAI at Scale: From Prototype to Production" - Powered by Zühlke

 

The presentation focuses on the challenges and solutions related to deploying and scaling Generative AI (GenAI) in financial services.

 

Current Landscape of GenAI in Financial Services:

  • 75% of financial institutions are using AI, but only 18% know how to scale and monitor large language models.
  • Significant organisational challenges exist in scaling GenAI solutions.

Challenges in Operationalising AI:

  • Many companies struggle with "PoC-hitis," where proof of concepts (POCs) fail to generate meaningful value.
  • Organisational capabilities like data governance, personnel, and technology are crucial for operationalisation.

Case Study - European Insurance Company:

  • Initial problems included long response times and the need for subject matter experts.
  • Expectations for fast resolutions and 24/7 availability were high.
  • Zühlke implemented a multi-agent chatbot architecture that improved operational efficiency and accuracy.

Results and Achievements:

  • 50% reduction in time spent searching for answers.
  • Accuracy over 95% with a significant reduction in error rates.
  • Net Promoter Score (NPS) of 80, indicating high user satisfaction.
  • Improved onboarding and reduced training time for new employees.

 

To successfully operationalise GenAI, organisations need to focus on user engagement, clear governance, and the alignment of technical solutions with business objectives.

 

Speakers:

  • Andrea Perl, Data & AI Lead, Zühlke
  • Shaun Lai, Data Lead APAC, Zühlke

Join the GFTN Network