At the outset, Pendo set into identifying the hurdles that tagged along with the data deluge impeding discovery and delivery of information for financial institutions. “The overwhelming challenge is reliability of data and the critical ability to act on it,” adds Alex Britnell, Sales Director, Pendo. Although CIOs see the potential of this abundance of data and appreciate the benefit it can provide to their decision making, inaccuracy of information is a considerable element of risk. Taking into account, the multiple, disparate sources that were involved in collecting data, such as stock exchange updates, investor portfolios and the dynamic nature of the financial industry, dubious looks are raised over the accountability of data. At its core, Pendo’s development team worked to provide a permanent fix to this predicament.
Where Data Becomes Information
“The flagship tool, Pendo Data Platform (PDP) was conceived to perform the required data analysis to deem data reliable and actionable,” Britnell says. “It shines the light on “dark data” and validates data relationships and lineage across all sources involved.” A synergy of data science, text analytics, and data management, PDP provides business users with the ability to classify, index, and search unstructured data across various file types, including PDF, Excel, emails, and text files. “Similarly, it also equips the Subject Matter Expert (SME) to traverse through structured data from within various databases and collate the files together,” states David Cohen, Chief Support and Quality Assurance Officer, Pendo.
Our data profiling algorithms capture a wide range of features from both semi-structured and structured data
Depending on the type of data, the platform utilizes specific plug-ins and mapping capabilities to locate the data. The core platform is underpinned by a powerful plug-in based architecture, where each plug-in represents a capability or algorithm that allows enterprises to handle different types of data structure, and efficiently analyze it.
Pendo banks on PDP’s key differentiating factors, such as its ability to be deployed and integrated with an organization’s existing infrastructure. “Unlike other products in the marketplace, our platform does not require a large technology footprint,” extols Zachary Mandell, Business Analyst, Pendo. “It can be positioned as a “side-car” to production systems, so as not to disrupt the current IT infrastructure.”
The function rich PDP platform is designed in such a way that its intuitive interface easily guides end-users through the process of data ingestion and discovery. It has the right ingredients—a portfolio of algorithms—to facilitate deeper data exploration and discovery. “For instance, our data profiling algorithms that capture a wide range of features from both semi-structured and structured data also support clustering and identifying patterns within sources,” says Philip Dodds, CTO, Pendo. Coupled with pattern matching algorithms, similarity rating for text and other formats, and the ability to generate a graph of relationships, PDP follows a cohesive workflow that allows financial institutions to extract valuable information from volumes of data. Users can also view the data within the platform prior to exporting, which assists them to accurately acquire data from documents and databases for use in other applications such as financial modeling.
Buoyed by the success of the platform in recent years, Pendo has boldly extended its arm of innovation into leveraging artificial and machine learning capabilities. “We are starting to focus on how these new technologies, combined with modern-day cloud storage and data analytical approaches can empower SMEs to capitalize on data,” explains Dodds.
Apart from machine learning, the company believes in continuous delivery, and is constantly scouting for ways to accelerate the path from data to value. “Internally, we push how new technologies, applied in innovative ways, can help the financial services benefit from the sea of data around them.”
To paint a picture on Pendo’s functioning in an active financial setting, Pamela narrates a recent customer success involving a large global bank that had a repository enveloping 23 million documents of multifarious types. In the process of feeding their risk model, the bank observed that a chunk of historical data pertaining to customer statements was missing. Stored in PDF formats of fifty different versions, the raw input consisted of position, transaction and summary records over a period of three years of historical data. Owing to the significant size of the repository and myriad nature of its contents, the bank knew that the situation requires a data specialist in the finance arena and Pendo entered the scene. PDP was deployed within a relatively lesser timeframe of 24 hours and the Pendo team began with the initial scanning of one million documents, so that they could understand and train the machine.
“By learning the patterns of data, we were able to configure and optimize the processing of the files, set filters and tolerance levels,” says Pamela. “Once the system was trained and tuned in the first four days through machine learning techniques, we started the bulk processing of the project.” At a remarkable rate of 600,000 statements a day, the Pendo team had completed the project—initially estimated to stretch through two years— within weeks, allowing their customer to reap sizeable benefits along with desirable ROI.
The Dynamic FinTech Industry
“The truth is,” Pamela says, “big data is infinite and indefinable—because as soon as you design a specific model, the data changes.” Especially with an industry as dynamic as FinTech, enterprises are seeking adaptable paradigms like that of PDP, which is flexible and scalable enough to embrace the change.
”Pendo Data Platform was conceived to perform the required data analysis to deem data reliable and actionable”
In an industry that invites data accumulation by the second—just like any other digitally transforming sector— there is a constant case of repurposing the data to raise the bar of comprehensibility. Even scanned paper-based data “deemed” digital, requires basic text analytics or image processing in order to be considered digital. From that, data digitization has come a long way with the influx of predictive analytics and cognitive computing, adding more value to data than ever before. Big data as a catchphrase will soon swing the spotlight it has been recently enjoying toward “fast data” and “actionable data.” When it comes to that, Pendo will be ready to brace the paradigm shift head-on by innovating better and more efficient ways of translating data into value. Driven by the common goal to help enterprises rapidly uncover value from data, the entire Pendo team with Pamela— the finance aficionado—at their vanguard point, believes that they have positioned the company’s stature ten months ahead of their expected roadmap.