How AI Automates an ‘Onerous’ Equity Research Process
Accelerating Equity Research by Automating Statement Analysis for Global Financial Service Provider
The Company
One of the global leaders in financial content services trusted worldwide in the context of tax, government, law, media and compliance. Providing the niche professionals with answers to complex problems, the company offers trusted insights, information and innovation to help unravel the knots in complicated situations in areas that move the world forward.
For the Chief Financial Officer at the company the areas of operation also focus on bringing efficiency and speed into the ecosystem of financial documents processing to make extracting and classifying key data simpler, helping the company take financial decisions based on accurate insights gleaned from it.
The Challenge
The lack of an incumbent to accurately process financial documents had significantly impacted the company in terms of accuracy, efficiency, data classification and extraction of key values from it. Data is rapidly rising to the top rungs of priority globally, and the need for getting it ready and handy drove the company to rejig their systems of financial documents processing. By automating aspects like document analyses, context building and classification, processing multi-page PDFs and accurate data extraction, the company wanted to create an environs around its financial data arm that helped the professionals weed out the onerous nature of the entire process while getting access to the required data simultaneously.
Results
Recognising the need to accelerate the process of financial documents processing, RecoSense identified the key areas for intelligent automation:
- Analysis of financial documents
- Identifying key financial concepts with a view to auto-classify extracted information and build context for quicker comprehension
- Imbuing the system with the capability to process documents with multiple pages in lesser time
- Accuracy in the extraction of crucial values from information-inundated documents
https://recosenselabs.com/wp-content/uploads/2022/02/Statement-Analyser.mp4
By automating the financial statement analyser systems, RecoSense established the first step for inculcating efficiency in the process cascade that made downstream aspects flow freely. With a view to helping the system build contexts with the extracted information, a cognitive engine was put in place that could work with any category of financial documents. Furthermore, the machine learning module deployed for the purpose was capable of identifying the right summary statements from the whole document to provide the professionals with the crux of the specific financial concepts contained therein.
RecoSense also deployed a meta engine to identify the core values from a document and extract them with high accuracy. A simple user interface with functionality for producing insights and reporting made data perusal more straightforward and comprehensible.
Impact
With the solutions thus deployed, RecoSense achieved industry benchmarks in certain processes.
- Accuracy to the tune of 97% was achieved, the highest among the 40+ vendors in the industry, assessed over a two-year period.
- The error rate was less than 2% in a dataset of over 600 processed documents.
- In successful identification of financial data points, the success rate was 100%.
- Scalability to the tune of 400,000 documents per year was achieved.
RecoSense is dedicated to helping professionals and agencies automate key areas where insights that can revolutionise development and growth can be derived.