Insurance Policy Recognition Solution
2021-09-28 13:27:441. Project background
In recent years, China's insurance industry is flourishing. While business is expanding rapidly, competition continues to intensify. The amount of insurance user data has increased sharply. The insurance policy business system can already scan paper insurance policies and store images on the server. The value information on the insurance policy is entered by the entry staff, and the information is stored on the server. However, with the increasing number of insurance policies, insurance companies need a large number of operators to enter insurance policies, which seriously affects work efficiency and expands personnel costs. Therefore, it is imperative to introduce OCR recognition technology into insurance industry in a timely manner.
The introduction of OCR recognition technology has opened a new page in the management model of the insurance industry and will surely lead the reform and progress of the entire insurance industry.
Sunshine Insurance Group is one of the seven largest domestic insurance groups and one of China's top 500 companies. Product services cover auto insurance, property insurance, life insurance, health insurance, accident insurance and asset management. Nearly 2,000 institutions spread across the country, leading the auto insurance flash claims service industry. This project is the centralized identification of the insurance policy images uploaded all over the country by the Life Insurance Department.
2. Function Analysis
Software structure
Software function 1. User management 2. Batch recognition of insurance policies 3. Re-recognition 4. Automatic detection of insurance policies 5. Single process recognition and submission 6. task acquisition 7. Vertical proofreading 8. Batch recognition image viewing 9. Workload statistics 10. Proofreading rate statistics 11. Workload report generation 12. Program interface design
The OCR image recognition function realizes the automatic recognition of 172 fields related to the numeric items and ticked options of the application form, accounting for 80% of the fields to be entered in the entire insurance form, effectively improving entry efficiency and saving entry costs. Since many companies in the market are promoting mobile insurance now, the business volume of outsourcing input companies is gradually declining, so their manpower and attention has gradually decreased in this field. The impact on us is that the input cost is gradually rising, and the service quality is falling steadily. Through the promotion of OCR technology, the filling and scanning of institutional insurance forms are gradually regulated, and achieve the goal of saving costs and reducing reliance on outsourcing input companies in the end. Through the promotion of automatic recognition of insurance policy, we gradually find problems, optimize system, accumulate experience. With the gradual standardization of institutional operations and the improvement of recognition efficiency, the OCR recognition technology is not only used in new contracts, but also in the centralized entry of preservation and claims, and the saving of entry costs is immeasurable. Comparison of efficiency after adopting OCR policy recognition system: Test standard: 10,000 insurance policies Efficiency of manual work: 8 hours of full-load work per person per day. According to the most skilled entry staff, if one sheet is calculated in 2 minutes, 240 sheets will be entered per day, 10,000 insurance policies require 1 person to complete 42 days. Efficiency after the adoption of the system: The recognition is completed on the server side without manual intervention and can be completed within 24 hours. At this time, the personnel only need time to proofread the recognition results. Because the recognition rate is highly reliable, and staff can use the collection of words to proofread. The characters are: 0-9 and ticked option. There are 11 characters in total, that is, only 11 character types need to be proofed if we use vertical proof reading for 10,000 insurance policies, so it can be completed in 1-2 days for 1 person to calculate 10,000 insurance policies, which can greatly improve work efficiency and save human costs. |
