AI Agent: A smart team member in Customer Service at Credit Agricole
We implemented an AI Agent for customer service team. It supports claim processing and after-sales customer care.
ABOUT THE CLIENT
Universal international bank
Crédit Agricole is a global financial institution with a significant presence in 47 countries worldwide.
As part of the Crédit Agricole Group, one of the world’s largest banking groups, it offers a wide range of financial services to individuals, businesses, and institutions. The Credit Agricole Group serves over 53 million customers.
Challenge
What problem was the Bank facing?
The Bank had an ambitious goal of increasing the number of active customers. However, this increase in customers would also lead to an increase in the number of complaints and reports, and the Bank did not want to expand the team responsible for handling them.
Long customer service times and lower customer satisfaction were the main problems we faced before implementing Deviniti’s solutions. A major challenge was finding mechanisms that would shorten customer service times while maintaining the high quality of the process.
challenge
Optimizing processes related to handling customer complaints
Crédit Agricole Bank Polska faced challenges in improving after-sales service efficiency:
Incorrect classification of incoming documents:
- Problem: Manual handling of a large volume of documents resulted in incorrect categorization of cases.
- Consequences: Delays in document processing, potential financial penalties from regulators for missing deadlines, and decreased customer satisfaction.
Lack of prioritization for responding to letters:
- Problem: The wide variety of customer issues and the need to assess the emotional tone of letters (e.g., urgent requests vs. complaints) made the prioritization process time-consuming.
- Consequences: Extended response times for requests and a higher risk of customer dissatisfaction due to delays in responses.
Time-consuming preparation of responses:
- Problem: Responses to letters, especially non-standard ones, took up to 2.5 times longer to prepare than responses to standard requests.
- Consequences: Employees had to spend a significant amount of time drafting responses, which prolonged the overall process, reduced work efficiency, and increased the likelihood of errors in the replies.
Repetitive manual tasks:
- Problem: Many activities, such as processing routine requests (e.g., withdrawing marketing consents), were handled manually.
- Consequences: Increased employee workload, risk of errors, and decreased operational efficiency.
Risk of cases getting stuck at various stages:
- Problem: Cases could get stuck at different stages of processing, for instance, due to a lack of clear priorities or incorrect classification.
- Consequences: Delays in processing requests, customer dissatisfaction, and the risk of missing response deadlines required by regulations.
solution
AI Agent
AI-Agent as support for the bank’s customer service team
AI Agent supports document analysis and classification processes, as well as generating ready-made responses to customer inquiries. It is integrated into the bank’s existing system, which handles all after-sales processes.
The main tasks of the AI Agent are:
- Reducing employee involvement in non-standard cases
- Structuring customer data based on documents
- Full automation of standard operations
Automating customer request processing
AI Agent modules
-
Extractor
Receiving and preliminary analysis of documents
- The system scans the document and analyzes key information: what is the purpose of the letter and what results does the sender expect?
- The extractor determines whether the letter was sent from an individual or an institution (e.g., UOKiK, KNF), identifies the customer by ID and determines which product or service the letter concerns
- The extractor recognizes the emotional tone and style of the document’s content, to later assign it an appropriate priority (a polite request, an aggressive complaint, or an official document is classified differently)
-
Classifier
Classification and handling of reports
- The system automatically registers customer reports and immediately forwards them to the appropriate team
- The classifier finds letters that have been sent to the wrong department and automatically transfers them to the correct department
- The classifier recognizes simple customer instructions, such as a request to withdraw marketing consent, and executes them independently
-
Generator
Support in composing the content of the response
- The generator helps the employee prepare a message to the customer, based on previously sent responses – standard and non-standard
- The generator suggests ready-made paragraphs that match the content of the letter, which the employee can insert with one click. It also leaves the possibility of entering their own text
- The system continuously checks the content, linguistic correctness and readability of the generated response
- The ready-to-send letter in PDF format – on a company template, with the employee’s signature – is generated with one click
Numbers
Results achieved by implementing AI Agent
756 hours per month saved by the customer service team for more complex tasks
Even 95% accuracy achieved by AI Agent in document classification
62% of certificate applications are handled solely by AI Agent
32% of account closure applications handled solely by AI Agent
50% reduction in the time required to process documents
40% of cases handled significantly faster, thanks to AI Agent support
Results
Key goals
Integrating AI-Agent into the customer service team has helped Credit Agricole Bank Polska achieve its key goals:
- Response times to customer inquiries have been greatly reduced.
- AI-Agent provides immediate responses for simple requests, while more complex cases are resolved up to 50% faster due to accurate inquiry classification.
- The collaboration between the customer service team and AI-Agent ensures responses are accurate, compliant with procedures, and more personalized.
- Faster inquiry processing ensures compliance with regulatory deadlines
Results
What our client says about the implementation
We wanted to combine ‘smart technology’ with the professionalism of our advisors. Automation of selected processes means that bank employees receive a ready-made semi-product and can devote more time to personalizing and empathizing with customer communication.
The proportions of time that employees spend on handling repetitive processes and the time they invest in personalizing communication have changed – to the benefit of our customers.
The prepared solutions were implemented primarily with the external client in mind and have proven to be excellent in this area. However, it is also important to remember about the new generation of employees who are more willing to use the latest technologies.
We notice the need to move away from established solutions and invest in next-generation tools to provide employees with job satisfaction and comfort.