Workflow Automation
The combination of advanced text processing and ML-based matching has increased the team productivity through reduction of routine operations, sparing the specialists from necessity of manual user request reviews.
Challenge
Our client has to deal with thousands of customers' requests via e-mail every week, so, taking the activity scales into account, the load on their support team is enormous.
They asked for a solution that would automatically classify these e-mail requests and extract vital information from the messages themselves. As a result, based on this analysis, the request should be transferred to the most suitable specialist, as well as the right equipment supplier should be quickly found.
About the Client
Our client is a large international distribution and retail company with 30 years of experience, about 1 000 goods providers and more than 2 000 employees.
Region
Europe
Time
9 month
Team
5 specialists
Solution
After the initial conversation we provided the client with our solution. Every customer claim, whether it’s a text message or an attached file, goes through advanced data processing based on ML and LLM algorithms.
First of all, the system extracts the most vital data pieces, including brand names, type of goods or some keywords that define the purpose of customer request.
All this happens thanks to text extraction, using the optical character recognition (OCR) technology whenever necessary, in the combination with LLM algorithms, which allow to identify the necessary text parts and classify the message according to its intention. As a result, it spares the specialists from manual email reviews.
When all the necessary information is extracted, and the claim intention is identified, it comes to automated service and product matching. The implemented ML model allows to compare the extracted data and customer intention to a company database or catalog, providing quick access to the most suitable goods and services provider.
It is worth noticing that matches can be both exact and partial and it mostly depends on how detailed the user claim is.
As a result, the analyzed information and selected matches can be transferred to the company specialists for decision making or further reviewing. The process itself is demonstrated in the following scheme.
Overall, the proposed solution speeds the company workflow up, reducing the number of manual operations and saving dozens of hours every week without productivity losses.
Want to obtain your own AI tool for better procurement
and customer support? Share your needs with us.
and customer support? Share your needs with us.
Key Features of Our Procurement Software
The solution we provided resulted in intelligent user queries management. Through advanced data processing and analysis our system significantly improved the work of the customer service department.
Data Classification and Extraction
Relying on Document AI techniques and LLM driven analysis, the system can automatically analyze every user query sent via e-mail. As document AI feature extracts key points from every message processed, including those presented in the form of plain text, PDF or spreadsheet, it allows to classify the claim, defining whether it’s a purchase, maintenance or support request.
Priority Categorization
Based on LLM driven analysis, the system also categorizes every user claim, giving it a special status regarding its urgency and complexity. In further perspective, the claim can be transferred to a particular team or specialist qualified enough to solve the problem.
ML-based Automated Product Matching
The usage of machine learning and natural language processing techniques is very useful when the task comes to choosing the right product or provider. Processed user queries are matched to the company database, which provides quick access to the necessary goods and services.
Key Results
The implementation of our solution leads to a few crucial benefits:
Better Retail and Supply Management
The company team received the opportunity to quickly access the most suitable good providers and transfer the user claim to the specialist needed, which raises the efficiency of support and management department work.
Real-time Analytics
Thanks to the advanced data processing, the procurement specialists are able to get the most vital information from user claims which increases the speed and quality of decision making and provides smooth communication.
Our Expert
"As a PhD expert who has dedicated decades of work to the AI and ML technologies, I can say that the future of procurement and customer support will certainly be based on wider usage of the solutions similar to ours. AI tools implementation leads to cost efficiency optimization and advanced decision making through reduction of manual operations and mistakes based on human factors."
Nikolai Karelin - Head of AI
Frequently Asked Questions
Not at all. You don’t need to have specified knowledge to implement AI software for your procurement solution. We are ready to provide you with all the necessary expertise, describing the process and project specifics in the simplest and most detailed way.
Indeed. When you see the profits of AI for procurement, you may seek for further advancement. We can provide you with updates and new features, for example, predictive analysis tools that can make decision making and product matching even more efficient and time saving.
Project info
Have a project in mind?
Reach out to us. We’ll make something awesome together.
