The CortexIntellect team developed an AI agent for a company from Austria that operates in the wholesale and retail trade of alcoholic beverages.
AI agent was created to automate the analysis of commercial offers from suppliers that come in the form of documents and letters. The AI agent processes large amounts of data, extracts key information about the product range, prices, and terms of cooperation, and helps to quickly compare offers with each other. This significantly reduces the time spent on processing documents and helps managers make more informed purchasing decisions.
Problem
The company deals with a large number of suppliers and regularly receives commercial offers for products. Trade offers come in various formats: emails, PDF documents, Excel files, images, or instant messages.
Before the system was implemented, procurement department employees had to manually open letters and messages and analyze commercial offers.
This created a number of problems:
- a large amount of manual labor;
- high probability of errors;
- lack of a single database of proposals;
- the complexity of analyzing price dynamics.
The lack of structured analytics made it difficult to determine the most profitable periods for purchasing goods.
Decision
What does an AI agent do:
- independently analyzes incoming messages and attached files;
- determines whether the document is a commercial offer, establishes its format;
- reads data about goods, services and prices;
- prepares structured information for further processing in the system;
- adds information to the company's internal CRM system. Each supplier's offer is linked to the corresponding product in the database;
- forms a history of offers for each product;
- builds analytical graphs of price changes.
Where is an AI agent used?
The AI agent works within the company's internal CRM system. It is integrated with communication channels such as:
- Email: The agent examines emails with subject lines and attached files.
- Messengers. The system processes supplier messages and attached documents. After receiving the notification, AI starts the process of analyzing and extracting data.
Basic functionality of an AI agent
Analysis of commercial offers
The agent is connected to the company's e-mails and messengers. The system analyzes all incoming messages and determines whether the received letter or document is a commercial offer or a regular message. If not, the agent ignores it. If a CP is received, the system automatically reads key data from the document: product name, price, sender, date and other important information. After that, the data is structured and transferred to the internal system for further processing.
Table of offers for each product
For each product, a table of supplier offers is stored in CRM. It displays the offer date, price, and company name or partner name. This allows you to see the offer history and compare counterparties.
Recognition of different file formats
AI is able to process various structures of attached documents: text messages, PDF documents, Excel spreadsheets, images, scans, photos. The system uses document processing and OCR technologies.
Integration with CRM product database
The CRM system stores the company's product database. After analyzing the commercial offer, the AI agent finds the appropriate sales item in the database, adds a new supplier offer, and updates the product offer history.
Automatic product matching
The digital assistant independently compares the products from the commercial offer with the products in the CRM database. If the system is not sure of the correctness of the comparison, a human verification process is launched. A purchasing employee confirms or corrects the found product.
Price analytics
The system builds price change graphs for each product. Analytics include minimum and maximum prices, offer dynamics, and average price.
Seasonality analysis
Based on the accumulated information, AI can determine the most profitable purchasing periods, seasonal price changes, and long-term trends in the cost of goods.
Customer journey
Step 1. The supplier sends the PO via email or messenger.
Step 2. The AI agent receives the notification and starts the analysis process.
Step 3. The system recognizes the content of the message and attached files.
Step 4. AI extracts data about products and prices.
Step 5. The system compares sales items with the CRM database.
Step 6. If the comparison is in doubt, the employee must confirm or replace the product.
Step 7. Supplier offers are added to the product offer table.
Step 8. The system updates statistics and price charts.
Step 9. The purchasing department uses analytics to make purchasing decisions.
Technology stack
- AI-based document processing → allows the system to process documents, highlight key data, and structure information for use in CRM.
- OCR text recognition → allows you to work with PDF files, photos and scans of sales offers.
- Natural language processing → helps the system understand the content of supplier messages, determine product names and prices.
- Database analytics → allow you to track price dynamics, generate statistics and analytical indicators.
- Cloud data processing → ensures system stability and scalability.
- Integration with CRM system → for automatically adding supplier offers and updating product cost history.
- Integration with email and instant messengers → to automatically receive supplier messages and attachments.
- OpenAI API → for studying the texts of price offers, recognizing information and organizing it.
- Python (CRM) → programming language in which the logic of the AI agent and integration with the CRM system are implemented.
Business value
- Automation of commercial proposal processing. AI analyzes incoming requests from suppliers without the participation of employees.
- Reduced workload on the purchasing department. Employees do not need to manually process documents and transfer data.
- Formation of a database of supplier offers. The company receives a systematic history of offers for each product.
- Procurement price analytics. The procurement department can analyze partner offers and price dynamics.
- Procurement optimization. The system allows you to determine the most profitable periods for purchasing goods.
- Increasing the efficiency of the procurement process. Purchasing decisions are made taking into account analytical data.
Implementation results
The implementation of the AI agent allowed to automate the processing of supplier CPs and integrate this process into the company's internal CRM system. The business received a centralized database of product offers and a purchase price analytics tool. As a result, the volume of manual work of the sales department was significantly reduced and the quality of trade decision-making improved.
Want to automate your commercial proposal processing and procurement analytics?
We'll analyze your processes, customize document processing, and implement an AI agent in your CRM to meet your business needs.
👉 Submit a request – let's discuss your project.
