XFintell – Semantic Solutions for Financial Intelligence

The XFintell product suite provides solutions for innovative collaborative business reporting and Deep Fundamental Analysis (using statistical machine learning and deep-learning approaches and semantic technologies like Ontology and Model Logic) that use knowledge about the reporting domain for which the reports are created. This new paradigm of writing and analyzing financial reports provides advanced real time guidance and validation support while the user is creating the business report.

The benefits of the XFintell knowledge based approach are:

  • The reporting process is under constant QA and errors are prevented from entering the process greatly reducing the overall costs of the reporting process.
  • The reporting process provides real-time collaboration that allows users to work on the latest shared and up to date version of the business report greatly improving team efficiency.
  • The reporting process uses semantics about the reporting domain to provide superior guidance during all stages of the writing process greatly improving report quality & efficacy of the team.

The paradigm shift on which the XFintell suite is based is the insight that if structured information is combined with the formalization of reporting domain semantics in a single solution, an entirely new and very powerful approach to internal and external business reporting opens up. No longer do we need to go through multiple cycles of writing free text, trying to make structured information from an unstructured input, running validations off-line and then hope all of this complies with regulatory reporting requirements.

By ensuring that the whole process begins and stays under control our customers can be confident that the reported information is always in compliance with the regulatory requirements – from start to finish – and costly rework can be prevented.

Some background

The knowledge about the reporting domain is captured in business rules and commonly used reporting patterns for the domain that align closely with how the domain experts and end users think about their information. For instance the features that support financial reporting have knowledge about reporting patterns such as roll-up patterns or roll-forward patterns and support for templates that provide correct and consistent sample statements and disclosures that ensure that the numbers cross and tie across tables and different sections in the report.

Using formalization of domain knowledge and the results of data analysis using machine learning algorithms on the corpus of over 40,000 financial statements in the SEC repository, allows the users to capture a deep understanding of reporting and disclosure requirements. For instance if – in the case of financial reporting – materiality thresholds are crossed the solution will detect that additional disclosures are needed and it provides templates for such disclosures to the user. It places the knowledge and experience of the best CPA as a permanent resource besides the end user of the reporting solution. For this feature the solution uses a powerful light weight expert system shell that runs in the browser and that pull relevant domain knowledge from a server library.

In other domains the challenge might be that multiple regulatory bodies each require different subsets of operational information. For instance different agencies might require different but sometimes overlapping environmental information. To manage these multiple reporting requirements is a complicated and costly matter. By capturing reporting requirements in rule bases a reporting process is capable of assembling the required business reports from the base information. Combined with co-authoring of text segments of the report this provides a unique blend of free text and structured data. This approach has been used to great effect in some government reporting solutions based named Standard Business Reporting (abbrev. SBR) solutions where the filing entities reported a set of information once and the solution assembled reports per regulatory body thus greatly reducing the reporting burden for the filing entities. A similar approach can be used in reverse if the reporting entity uses the domain semantics and rule bases of the XFintell solution to collect information once and report to many different regulatory bodies as needed.

Besides providing features to manage the “understanding” of the respective reporting domain, the XFintell products provide cloud-based real-time collaborative authoring of business reports. Team members can write in the same report at the same time and the solution shows in real time the action results and locations of other users’ activities in real time. No longer is there a need for juggling of parts of documents and doing expensive merging and synchronization activities. Everyone in the team works on the most up to date version of the document and its consolidated information. Access control is very granular preventing overlap of activities on similar sections in the report. This greatly improves the efficiency of the teams while authoring business reports. Access to statements and disclosures to view out change information be managed based on organizational authorization scenes and roles.

If all of this spikes your interest drop us a note at info@xfintell.com.

Leave a Reply