A Fortune 50 bank, with a strong presence in institutional, commercial, and retail banking across more than 160 countries and jurisdictions, knew that the way its agents handled customer queries during calls needed to change. Its Personal Banking and Wealth Management (PBWM) division was working with tools that offered only limited contextual suggestions and simply couldn’t scale personalized insight at the pace the business needed. That gap showed up everywhere it mattered most; in agent efficiency, in response accuracy, and ultimately in the overall customer experience. The bank saw an opportunity to close that gap with AI, but doing so safely, at the scale of a Fortune 50 institution, meant it first needed a scalable framework to evaluate and implement multiple AI models securely and compliantly and not just apply AI, but adopt it the right way.
Ascendion designed and implemented a GenAI studio to evaluate and compare multiple LLMs in real time using real call transcripts, built around a model-agnostic pipeline that ingests call transcripts, extracts context, and surfaces intelligent prompts to support agents in the moment.
A prompt engineering framework was created to manage and tune context-specific prompts across various models, all running inside a secure, compliance-friendly sandbox where the bank could safely run model experiments and monitor performance across response accuracy, latency, and hallucinations. From there, Ascendion developed 20+ use cases spanning call summaries, transfer detection, knowledge extraction, and guided response generation, all powered by an Agentic Framework with built-in guardrails, feedback loops, and continuous improvement through real user testing.
Tech Stack:
Delivered up to 3x faster agent response time through real-time GenAI prompts for a Fortune 50 bank.
Improved customer satisfaction scores for 10M+ by delivering context-aware, accurate replies.
Created a secure GenAI experimentation environment, allowing safe exploration across multiple models.
Enabled scalability of agent assist tools across business units and markets.
Established a repeatable GenAI integration model, setting the foundation for enterprise-wide GenAI adoption.
