Summary
Key results
Average 85% reduction in manual work across targeted areas
About 92% improvement in the quality and timeliness of data delivered by DCN
Manual document analysis introduces risk of errors and delays
DCN needed to efficiently process massive volumes of incoming data from diverse sources – the quality and freshness of which directly determine the value of its services. Manually handling hundreds of thousands of documents each month slowed teams down, increased error risk, and hindered rapid delivery of critical market insights to clients. The company required a scalable, automated solution to ensure high data quality, accelerate processing, and enable consistent, reliable analytics.
AI-powered data extraction and cloud integration
Sii’s AI experts, working closely with DCN, identified the most impactful optimization areas and implemented solutions on AWS using Amazon SageMaker. The work combined initial document classification with standard methods and R&D initiatives for extracting and clustering information from complex construction projects, plus data deduplication using Sii-designed neural networks.
The scope included:
- identification and prioritization of high-impact optimization areas
- implementation and deployment of multiple AI-based solutions
- document classification achieving ~98% accuracy
- information extraction and clustering that cut analysts’ time by 85%
- deduplication and matching of new data to existing records for updates, automating ~92% of the overall process
- ongoing monitoring and maintenance of deployed solutions in production
Scalability, speed, and data quality
The solution significantly reduced costs and boosted end-to-end process efficiency by relieving staff of manual document review and data entry. Teams can now focus on expert tasks that directly drive revenue – such as engaging key customers and delivering accurate, rapidly accessible insights on projects of interest. DCN’s clients gain a competitive advantage by discovering new investments in their regions earlier than before.