Is access to more and complex data slowing productivity?

    XpertRule VP, Deepak Karwal, believes it is and there are ways to stop it.


    In a recent conversation with a potential customer, they highlighted that their mortgage underwriters had seen a drop in applications processed from over 6 to an anaemic 2-3 applications in a typical working day. Whether this is representative of the wider US mortgage market remains to be seen, however, this observation does start to shed light on an emerging challenge across industry sectors – is the ongoing growth of data actually hampering productivity?

    It is no surprise that the growth of data looks set to continue its acceleration, from close to 97 zettabytes in 2022 to a forecasted 181 ZB’s by 20251 (a single zettabyte is equivalent to 1021 or equivalent to 250 billion DVD’s, in other words, a lot of data!). In the ongoing shift to digital, organizations have committed significant resources to harvesting, storing, and analysing this data. Fuelled by mantras of ‘data as the new oil’ organizations have invested in creating these vast potential fortunes, vast ‘data oil fields’ in the form of data lakes, ponds and would you believe it data puddles, in the belief that within them lies hidden the future source of long-term growth for the organization.

    For most enterprises much of the efforts invested over the last several years have focused on using the data for reporting and Business Intelligence capabilities, and to far lesser extent on building predictive AI models.. What most organizations have not succeeded in doing is to operationalize the use of their vast data to transform their businesses. Without operationalizing the data, organizations cannot achieve a return on their investment in harvesting the data and worse still rather than improve operations this data can turn into information overload slowing down critical decision makers such as mortgage underwriters.

    There is now an emerging set of technologies collectively known as Decision Intelligence that can help organizations operationalize their data by turning data insights into automated actions. Gartner define Decision Intelligence as ‘a practical discipline used to improve decision making by explicitly understanding and engineering how decisions are made and how outcomes are evaluated, managed and improved by feedback’. The two key components of Decision Intelligence are Decision Automation that can capture and automate the decision-making expertise of subject matter experts and Machine Learning for discovering new decisions and/or actions based on predictive patterns from data.

    The emerging decisioning technology provides a new dimension to automating business actions. For example, the technology can use machine learning to suggest the correct promotional discount to prevent customer churn or a ‘zero-touch’ approval of a low value claim on a home insurance policy. By importantly also allowing a decision feedback loop that identifies the success (or failure) of the action taken, the impact of those decisions can be understood by the business and refined to align to the right business outcomes. Decisioning makes micro-segmentation, personalised relationships, and individualised customizations a reality—it puts your data to work.

    Decisioning technologies are emerging with the capabilities to integrate event information with enterprise customer transactional data, that is enriched through third-party bureau data, superimposed with a predictive model and directed with pre-defined automated decisioning expertise and business rules. The outcome is a decision, one that can be fully automated or a decision which can be augmented. This provides us with the capability to leverage the power, predictiveness and rationality of the machine with the cognitive eloquence, emotional perspective, domain expertise, and contextualisation only a human can provide. These decisioning technologies can also direct decisions like a Beethoven symphony, effortlessly orchestrating decisions between humans and machine at a digital scale supported by cloud computing horsepower. For example, automating the selection of the right supplier based on complex combinations of transportation environmental impact and contractual volume discounts, to bringing humans ‘into the loop’ when a customer’s request for medical approval of a simple treatment indicates potentially underlying medical concerns given their complex history.

    In the burgeoning mortgage industry, data sources to support underwriting have seen rapid growth, from evolutions to traditional customer credit score information to localized specialist third-party bureau data services to data aggregators that use customers utility records and bank data to create cashflow ratios. Trying to manually process and interpret this level of data availability has exceeded the capabilities of humans. Even some of level of decision support, where relevant data for the decision may simply have too many inferences, permutations, and combinations for a rational decision to be the outcome. Even with the highest performing underwriters, this plethora of information, together with the challenge of ever-changing demands on assessment criteria, it is no surprise that underwriter performance is being impacted. While the overall decision for a mortgage loan may be better, the likelihood of not identifying key loan risk correlations hidden in the complexities of the data is high while overall productivity is significantly impacted. This trade-off between quality and quantity can effectively be minimised through decisioning technology.

    As data continues to grow exponentially in parallel to a maturing across the AI lifecycle, opportunities to drive a step change in the customer experience, productivity, and growth by applying data insight and decision automation have never been so high. The emerging discipline of decision intelligence provides the capability to democratise actions from data at a scale never seen before transforming data blindness to 20/20 execution clarity.

    Deepak Karwal is a vice president at XpertRule, a leading provider of digitally intelligent decisioning solutions and author of ‘The Automated Enterprise: Digital Reinvention through Intelligent Automation’.

    (1) IDC’s Global DataSphere Forecast