Insurance.
Use case for omnichannel conversational AI for inbound customer service.
The transformation of first notification of loss (FNOL) processes for motor glass claims enabled an insurance provider to improve the speed and efficiency of inbound customer service through an omnichannel conversational AI solution. By integrating voice, WhatsApp, CRM, policy, and claims systems, the digital agent was able to manage high volumes of customer interactions seamlessly, delivering faster resolutions, improved accessibility, and a more connected customer experience across channels.
Problem statement.
The client’s SOS and Claims line was experiencing high call volumes for motor vehicle glass claims. The process was manual, repetitive, and required extensive agent time for policy validation, eligibility checks, and vendor location. The client needed a scalable, automated solution that could handle direct personal-line customers while maintaining high service quality.
Solution overview.
Connect deployed a conversational AI agent to automate the end-to-end FNOL process for eligible motor glass claims. The solution was designed to:
- Handle direct personal-line customers with a single active policy.
- Automate policy validation and eligibility checks via a custom middleware layer, abstracting logic from backend systems.
- Manage exception handling by transferring ineligible or complex cases to human agents with full context.
Key functionalities.
1. Policy lookup and validation
Middleware checks for policy status, cover type, payment status, and incident date alignment.
2. Dynamic exception handling
Agent transfer triggered by specific “agent transfer reasons” with contextual metadata passed to human agents.
3. Extraction
Extraction of numeric identity (ID) numbers and alphanumeric data (e.g. vehicle registrations).
4. Confirming location to determine the nearest repair vendor:
- Primary: Confirm address from user records.
- Secondary: User-provided postal code (converted to latitude/longitude).
- Tertiary: Whatsapp sent to the customer to capture their vehicle location and send back to the digital agent, all while on remaining on the telephone call.
The outcome.
Automating 6 minute first notice of loss
calls end to end.
Improved customer experience, surpassing the target
Net Promoter Score (NPS) by 8%.
Managing noisy, incomplete and delayed data.
The solution was designed to be resilient to delayed data, which are common challenges in real-time customer interactions.
- Noisy data: For voice inputs, we implemented audio signal processing and custom ASR models trained on diverse acoustic conditions to handle background noise, accents, and poor line quality.
- Incomplete data: The conversational flow was built with adaptive logic and graceful degradation strategies. For example, if a customer provided partial ID information or OTP delivery failed, the agent would prompt for missing details.
- Delayed data: Backend latency was mitigated through asynchronous data retrieval and caching. The agent could continue the conversation while waiting for system responses, maintaining engagement and reducing perceived wait times.
Watch how Elerian transforms customer interaction with advanced AI:
About'Connect.
Connect combines global contact centre and customer experience (CX) expertise, deep domain knowledge, and unparalleled industry skills to make the complex, simple. Since 1990, we have leveraged our vendor-independent managed services approach to digitally transform how organisations communicate, both internally and externally. We specialise in combining the most relevant technologies and services from leading vendors and platform providers to create opti-channel engagement solutions, orchestrating frictionless experiences and simplifying complex communication challenges.
Connect with us Connect UK, Connect South Africa, Connect India, Connect USA.
Find out how we can help your business communicate better.
To discuss your communications challenges and requirements, get in touch with us today.
Connect with us now.
New web: Contact Us
"*" indicates required fields