AiThority Interview with Anthony Scriffignano, Ph.D., SVP and Chief Data Scientist at Dun & Bradstreet

Can you discuss the emerging trends in data integration and data transformation that expose vulnerabilities in enterprise data management frameworks?

“Within the enterprise sector, data integration and transformation has shifted the overall management framework in various ways. For instance, consider a federation of capabilities where there is a single database connection to access and use a broader cross-section of data that is in the cloud, on premises, or a combination of both. This trend is not only being propelled by the tools available and the corpora of data, but also by the rising expectations we’re seeing from businesses to draw on datasets that leverage new types of insights to make informed decisions.

“In fact, the dependency on data is accelerating faster than ever before due to the global disruption caused by the pandemic which has, in turn, caused supply chain perturbations. However, this disruption is not diminishing customer expectations, where there’s increased demand for agility and resiliency when it comes to goods and services. Those who discover, curate and make sense of data are under increased pressure – as they now have to work at an unprecedented pace to meet ever-growing customer expectations. In many cases, the change in the data is slower than the change in the environment, so new skills related to understanding biases and deficiencies within datasets is increasingly coming into focus. 

“Similarly, further external factors are also altering data management frameworks as we’re experiencing increased governance – with an intense focus on how data is sourced and used around the world. This shift in regulation is creating an important dialogue, with discussion around the difference between what ‘can’ be done and what ‘should’ be done within the data and analytics industry.”

Your latest report” Future of data” points to struggles within organizations related to data management and fraud. Why do you think CIOs are lagging behind in meeting business goals linked with data management? 

“Instead of seeing it as a lag, another perspective on why many CIOs are increasingly behind when it comes to meeting business goals linked with data management is due to an exponential increase in expectations with regard to datasets that need to be curated; especially because of the turbulent economic landscape. This “more and faster” culture as a result has led to the skills that may have once driven success to no longer being sufficient for the future.  

“Elsewhere, as fraudsters and other malefactors do not abide by regulatory confinements when it comes to how data is used, they can perpetuate new types of fraud, such as ‘deep fakes,’ which is when videos, images or audio recordings have been manipulated by AI technology.

We have acquired different definitions for data literacy. As a data scientist, how do you define data literacy and how does it influence the working hard skills required to handle big data, computing and analytics?

“Data literacy can be defined as the basic understanding and competency when working with all facets of data; the discovery, curation, synthesis (analytics and sensemaking), fabrication of results (storytelling/making the conclusions useful) as well as governance and quality assurance. A crucial element of data literacy is having a holistic view and understanding the importance of interplay e.g., poor curation can result in inappropriate samples of data, leading the AI to draw false conclusions.”

What kind of skills should a data-driven company hire for in 2022? Despite so much supply of data science professionals in the recent, why is there a serious gap in the hiring industry?

“In terms of skill, it’s important the look at the broad set of capabilities outlined above. As the Chief Data Scientist at Dun & Bradstreet, I look for someone who exhibits scientific thinking – they must be able to devise well-informed questions, complete the appropriate research, select the best analytics methods, analyse and understand biases that may impact conclusions. It is also important that they can collaborate well with others, as the nature of many contemporary problems requires critical sharing of ideas and capabilities.

“Alongside the professional attributes needed to be a data scientist, the hiring gap is also becoming increasingly complex due to expectations around job mobility, increased capabilities in technology which demand new and different skills, and social/endemic changes driving some workers into less permanent roles.”

Your take on the emerging trends in AI Machine Learning influencing data management and security analytics platforms? 

 “The emergence of trends in AI machine learning can be broken down into four categories: 

1. The technology: As AI continues to mature in enterprise applications, there’s an increased focus on what technology is used. Currently, there is a plethora of open-source and commercial software available which can do most of the heavy lifting, however, managing the selection and subsequent effectiveness of toolkits on the market is no small undertaking. Developing the remaining technology is no small task, especially in the context of changing regulations and evolving technology landscape. 
2. AI cannot be seen as a single component as it has many parts: There is a myriad of methods when it comes to implementing AI, such as supervised, unsupervised, cognitive and hybrid methodologies. Therefore, it’s extremely important to select a process for the right reason, not simply for convenience or because one method is preferred by practitioners. For example, conclusions reached from one method of AI may be vastly different from those produced by an alternative process. Now more than ever, it is imperative that methods and processes are considered with a clear purpose in mind and constantly re-evaluated as conditions and available capabilities evolve
3. A humanistic element must play a part in emerging AI trends: Currently, emerging capabilities must not only consider the pre-existing workforce but the incoming workforce. Constant re-education is not a ‘nice-to-have’ in this field, but rather a must. 
4. Having the right mindset is important when it comes to emerging trends, and it’s often one that is overlooked: Various mindsets are critical, including empirical rigour (being able to construct methods that are clear and reproducible), the importance of analysing sampling and pinpointing biases, along with the understanding of data cleansing and transformation. Effective leaders in the sector have a handle on the specific epistemologies, belief systems and best practices – with an eye on developing operations in accordance to the changing landscape.”

Your advice to CIOs and data officers on how to priories data planning and integration? Which top tools would you recommend for meeting data integration goals? 

“My general advice would be to make these efforts a team sport. The field is simply too complex to do it alone. It’s crucial to have the right partners when it comes to data acquisition, sensemaking, management and transformation. It’s equally critical to make these relationships strategic and mutually beneficial to drive long-term success.”

Thank you, Anthony! That was fun and we hope to see you back on AiThority.com soon.