Why should Big-Data and AI intersect with business workflows?

Ferris-Matt-hew Why should Big-Data and AI intersect with business workflows?
Matthew Ferris, Vice President of Analytics, Decentrix, Inc

Despite all the Big Data and AI discussions that focus on very niche business application, very little of this technology has found its way into the workflow systems that actually run the myriad business workflows, processes, and decision making on a daily basis.


Consider, the following

  • Most workflow systems produce siloed reports, based on specific operational requirements, and generally require human interpretation and subsequent action. It could be argued that these reports actually create more work in the form of human interpretation needed to create actionable outcomes.
  • In many organizations, Big Data is largely an augmentation to, or byproduct of, workflow systems. Transactional system reports cannot accurately reflect the breadth of large data repositories and big data research not founded in deep business knowledge is simply whitewashed with beautiful, yet meaningless graphic renderings.
  • If the data is not accurate, or at least reflective of the actual business, present or historic, how can any AI engine possibly achieve any degree of insight or decision‑making acuity for the future?

We observe in most businesses, a disconnect between the business workflows, decision-making processes, and the promises big-data and AI will provide. It is this disconnect that invariably determines the success of projects.


The case for big-data and AI have been largely impeded by four primary actors. Their perspectives and motivations, and limitations differ, and without exception, their language disparities create chasms of comprehension, invariably leading to less than optimal outcomes.  Classic technology vendors, consultants, workflow vendors, and IT organizations are motivated by a drive for market-share, head-hours, sustainable franchises, or suffer from a lack of development capacity.  Clearly, the specific needs of the business are second to the motivations and limitations of each of these actors. While rare, the key is finding an organization that understands the business, has the technology, process expertise, and the capacity to build integrated solutions.


The hard work starts with the organization and skills necessary to deploy such a transformation.  At the heart of this organization is a culture of discovery that that integrates data engineering teams that can ensure quality data, data science teams that can interpret data and model alternatives, business experts that ensure data reflects the reality of the business, and application teams that can integrate with existing processes and build new apps that effectively surface the intelligence in an easy to use way.

Once in place, the only way to tackle a data-driven transformation challenge is to focus on the key revenue generation streams in the business, but with a customer-centric view. Customers are the mirrors of the marketplace’s soul and it is essential to consider their interactions with your business as an integral part of the supply chain.

This lays a very different foundation for big data.  Capturing data along the entire supply chain enables an enterprise to start analyzing strengths, weaknesses, evaluate metrics, understand business behaviors, and explore customer behaviors.  This creates the foundation for an AI strategy of distributing intelligence to different parts of the workflow and concentrating knowledge and learnings for maximum corporate effectiveness.

This approach is not for the feint hearted. This is where all our actors invariably fail. Everyone strives for a quick win. but, there is no substitute for hard work.  It takes discipline to apply a methodical and scientific approach that challenges gut-feel precepts of the business, letting real-world data and metrics guide decisions and ultimately shape the culture.


Businesses are facing info-stress. They are collecting lots of data and spinning it around on rusty platters, paying for the privilege and hoping for a future where someone will turn that data lead into gold. The trick is to enable the right information, at the right time, to make and automate the right strategic and tactical decisions. Many enterprises are using institutionalized workflow systems, plugging gaps with big-data and AI, hoping that somehow this amalgam will buy them time, it can’t.

If we’re to be serious about the power of data in transforming business practices, we need to accept that existing workflow systems will not change at the speed and dexterity required, to face changing market forces.

The very act of transforming a business using the same tool that institutionalizes existing practices makes no sense. Equally, the very act of trying to transform business using tools that have no deep comprehension of the business and training this tool on inaccurate data is folly.

Any big-data initiative that is not intimately connected to your business workflow will result in failed AI initiatives. Deep process integration is key to process transformation. Augmentation is simply feel-good designer baggage.