Once every decade or so, we witness the birth of a new class of applications. We are on the cusp of the era of “Systems of Intelligence.” Systems of Record such as ERP and Systems of Engagement ( i.e. those that provide a user interface as accessible as consumer Web services such as social media) automate or simplify key processes, often by enabling access by a new class of users.
Systems of Intelligence take a geometric leap beyond the other two. Systems of Intelligence can deliver measurable and lasting competitive differentiation. Their power is in mining the engagement in real-time to anticipate, influence, and optimize customer experiences. In their most advanced form, these systems learn both from the user’s interactions as well as external observations such as an extended network of relationships. By continually improving themselves in this way, they finally bring (Gordon) Moore’s Law to software.
Line of business executives: More than ever before, line of business executives need a greater appreciation of the connections between these new applications and business strategy in order to be effective change agents within their organizations.
IT executives: The proliferation of new applications beyond those traditionally controlled by IT departments does not relinquish IT executives of responsibility for the success of Systems of Intelligence. Rather, they will be a strategic partner to line of business executives in helping to build or manage the radically new infrastructure that will support these systems.
Looking back in order to look forward
Like previous generations, Systems of Intelligence don’t replace the applications that came before but instead build on them. And each generation remakes and is also remade by the supporting software and hardware infrastructure. In this Professional Alert, we look at the lessons the Systems of Intelligence pioneers can confer on mainstream enterprises.
Before we look forward, however, we need to set some context. While Systems of Record don’t deliver sustainable competitive differentiation, they do deliver functionality on which Systems of Intelligence can build. So we start with a recap of Systems of Record and then show how Systems of Intelligence grow out of the more well-known Systems of Engagement that Geoffrey Moore describes.
Systems of Record: Foundation for future systems
These applications were mostly about internal efficiency. They saved money by standardizing business processes that are listed in P&L statements as Sales, General & Administrative so they could be easily shared across lines of business. They also encompassed some industry-specific processes, mostly in manufacturing. Capabilities such as taking and processing a customer order all the way to fulfillment and payment became everyday functionality.
By standardizing these business processes, the applications also standardized the transaction data. This data in turn supported analysis of historical corporate performance, such as sales by product by month. Since this analysis was historical, these systems were more limited in forward looking planning. You could think of them as helping to steer a ship by looking backwards at its wake.
Having a packaged software application that can manage all the steps from order to fulfillment with real-time data on products, pricing, and inventory can simplify the path to Systems of Intelligence. When Apple introduced the iTunes music store in 2003, they actually built it on top of their SAP R/3 system because they didn’t have the time or money to build the common functionality from scratch. This was a System of Engagement, which is a System of Intelligence without the analytics or machine learning that makes it possible to anticipate and influence customer interactions.
Apple has surely rebuilt the iTunes services over time on a more optimized platform. Similarly, Netflix, Amazon, and even Salesforce.com all started moving some of their services from traditional enterprise technology such as Oracle databases to Internet-centric, open source alternatives. But that enterprise technology was a critical platform upon which they created their initial Systems of Engagement.
Systems of Engagement: Consumer Internet services inform how B2C enterprise systems are built
Facebook, Twitter, LinkedIn, and the newer services such as Pinterest and Snapchat are all about giving consumers control of an immersive experience where they are connected to others. Google and Amazon are left out of this category on purpose and we will explain why later. For enterprise customers, these consumer Internet services provide examples of how to build a new generation of B2C applications. The first ones are appearing in multi-channel businesses such as banking and retail. They let the consumer take control of their individual experience interacting across channels of distribution and the online touch points within those channels.
Unlike Systems of Record, there is nothing standard about Systems of Engagement. They don’t come as packaged software, either on-premise or delivered as-a-service. Nor do they start with business processes, even the customer-facing ones. Rather, they start with data – about the consumer. They start with a coherent picture about each consumer that carries across channels and touchpoints. Then they require a lot of custom development. Systems of Engagement are most definitely a source of competitive advantage to those that implement them effectively.
Just like Systems of Record drove deployment of a new software infrastructure built around SQL DBMS’s, Systems of Engagement are remaking that foundation. In fact, more innovation is going on in data management right now than any time in over three decades, if ever. Even if line of business executives are driving digital transformation with these systems, IT executives will still be responsible for deploying and operating this new infrastructure. The next professional alert will go into this topic.
B2C Systems of Intelligence: Amazon, Google, Harrah’s Casino build on Systems of Engagement
The first Systems of Intelligence seem likely to be an evolutionary extension of Systems of Engagement, similar to their early transition from Systems of Record. Building on Systems of Engagement, Systems of Intelligence use machine learning to continuously update a model of customers so they can anticipate, influence, and optimize interactions in real-time.
The reason Amazon and Google belong here and not in the Systems of Engagement category is the intelligence they have always had built-in. Amazon was continually learning about each consumer from their every interaction. That learning drove ever more intelligent merchandising decisions via their recommendation engines. Google’s search results were also informed by user interactions. The ordering in the display of search results was continually refined based on learning which links users clicked on in addition to the raw relevance ranking.
Harrah’s Casinos has been accumulating more knowledge about every customer for almost 20 years. They now use loyalty cards to reach customers in real-time across every activity at their resorts. They know enough about each customer’s predilections to keep a high roller having a bad evening from leaving by offering her free tickets to a show with her favorite performer the next evening.
The intelligence in these systems comes from using the breadcrumbs tracking user behavior to feed a machine learning system that creates a predictive model of customers. That predictive model has enough intelligence to drive loyalty and profitability at the level of the individual consumer in real-time. Underneath that intelligence, Systems of Intelligence will likely share a similar data management platform with Systems of Engagement.
There’s a myth that Systems of Intelligence are suitable only for consumer applications. The Internet of Things needs the same type of intelligence but with far more numerous devices and sources on streaming data.
The management systems that run these incredibly sophisticated online applications for continuous availability have a predictive model of how the application should run. When the application starts acting outside the expected boundaries, the model suggests the best course of action to an administrator, if it doesn’t make changes itself.
The implications of these transitions will be profound.
Systems of Intelligence will support entirely new business models and methods of achieving competitive advantage. That will require a culture that pushes responsibility further out in the organization for pursuing new business initiatives and for implementing the supporting technology.
On the technology side, the core foundation is more focused on the data about the user or customer than the data management software itself. And that data management software is more likely to be a platform of tightly integrated products than a single one. More iterative delivery models will be needed to support the continual flow of application updates. Enabling that flow of updates will also require new application architectures based on micro services that can be upgraded more independently rather than the more monolithic legacy applications.
Wikibon big data research agenda
The next Wikibon Professional Alert in this series will explain in more detail how IT executives should consider the trade-offs inherent in choosing a data management platform, whether Oracle in one direction; AWS, Microsoft Azure, and Google Cloud Platform in another; or the Hadoop and Spark ecosystems in a third direction.
More broadly, the path from Systems of Record to Systems of Intelligence will be a major focus of Wikibon’s Big Data sector research. The point of view will be to explain how mainstream Global 2000 and mid-size organizations can learn from these applications’ pioneers in navigating these transitions.
Line of business executives charged with digital transformation must understand the dynamics of Systems of Intelligence well-enough to be effective sponsors of new innovation initiatives.
IT executives must be effective partners and understand how to build Systems of Intelligence on top of Systems of Record, blending in analytics from systems of engagement. They will also need to understand how to build a radically new infrastructure to support these systems.