La Poste: Using NLP to boost call center performance

In 2021, La Poste Solutions Business noticed a significant loss of customer information between the end of the call handled by its sales representatives and the addition of that conversation to its CRM. It therefore approached Lettria to benefit from its natural language processing technology and automatically enrich its CRM.

Project challenges

The contribution of AI and the NLP solution developed by Lettria had to meet several objectives:

• Limit the risks of compartmentalization and loss of customer information.
• Increase sales performance and the quality of customer relationships.
• Provide precise, operational performance indicators.
• Identify themes and trends arising from customer requests.
• Gain a global view of customer expectations.

To meet these needs, the aim was to extract key information from each customer conversation: a major challenge, given that La Poste Solutions Business handles over 1 million calls every year.

‍

The methodology

Lettria's methodology is characterized by an hands-on personalized approach. The aim is to obtain feedback from all users to create a tailored solution with them, in this case perfectly adapted to the context of La Poste Solutions Business.

Over the course of 4 months beginning in September 2021, teams from Lettria and La Poste conducted 2-hour joint workshops weekly to define the best approach and concretely identify the conversational data likely to be automatically populated into the CRM.

32 "Voice2CRM" labels were defined to qualify conversations. These labels were used to train the AI with NLP technology developed by Lettria, coupled with an API that automatically integrates the data into the appropriate CRM fields. Within 4 months, La Poste Solutions Business had an operational proof of concept.

Nathalie Crauet, Digital Project & Voice Program Manager for La Poste Solutions Business, notes that "The relationship between sales representatives and their customers is at the heart of what's at stake. Lettria's Voice2CRM tool shows very objectively the points and questions raised, and highlights the business actions that need to be improved. In the end, it's all positive, and helps us to improve our sales methods."

She also highlights the potential to expand on the collaboration thanks to the results of this first collaboration.

"Lettria communicates with our CRM on 24 types of data, structured and labeled. So there's plenty of potential for development. Marketing will be able to exploit the data and propose ever more bankable campaigns to our sales people."

‍

Project milestones

The implementation of Lettria's NLP technology in La Poste Solutions Business call centers involved 5 key stages:

1. Definition of a work plan based on a precise strategy and objectives, calibrated to La Poste's needs.
2. A development phase dedicated to integrating the platform and customizing the algorithms to ensure proper integration of the technology and optimize machine learning performance.
3. A test phase to evaluate the solution's ability to identify key information and automate CRM data entry.
4. Deployment of the solution on 100% of calls made in La Poste Solutions Business call centers.
5. Ongoing optimization based on user feedback and close collaboration with La Poste teams.

‍

Results and benefits

Before Lettria's intervention, only 3% of calls to La Poste Solutions Business were available for conversational analysis; now, 100% of calls are processed and key information is automatically entered into the CRM system. Thanks to the NLP technology developed by Lettria, La Poste Solutions Business has seen a 15% increase in call center performance in just 6 months, as well as a 50% reduction in the time spent manually entering data into the CRM. This all has meant a significant reduction in costs associated with CRM management and inbound call monitoring.