The enviable goal of the data-driven enterprise is a long-standing one, but it is continually frustrated by the challenge of getting data insights into the hands of not just the executive group but the whole enterprise. As much as it might disconcert some senior executives, much of the enterprises they lead are defined by the culmination of many small decisions on a day-to-day basis, not grand strategic visions. The infusion of artificial intelligence (AI) into an ever-growing range of enterprise applications is the "bottom-up" approach to building the data-driven enterprise, tackling persistent skills gaps, but the importance of a "top-down" approach, including enterprise-wide data governance, must not be forgotten.
It is hard to find someone who fundamentally disagrees with the benefits of making an enterprise data-driven. These benefits are largely rooted in the fact that enterprises are impacted by the swarm effect of countless small decisions made by people across the organization every day. Whether it is a customer service rep who better understands a customer's purchase history and makes a sale as a result, or the production line engineer making choices about which machines to use that day – these decisions are ultimately delegated by senior executives, the group that has typically benefitted from analytics. Business intelligence (BI) and more recently self-service analytics were meant to realize this opportunity, but while they did spread more information to more people, they have rarely infiltrated the whole organization.
Another approach, embedded analytics, has made significant gains, providing toolkits for developers (or the application vendors themselves) to put analyses in applications such as enterprise resource planning (ERP) and customer relationship management (CRM) – the "go to" apps many enterprise employees use day to day. This approach offers a means to tackle the challenge of getting analyses in front of many more people. It is AI that will lift these embedded analytical capabilities to a more widely useable form.
AI technologies tackle the remaining gap in putting more data-driven insight into the everyday of enterprises: data skills. Most analytics tools require training to use and analytical knowledge to ask the right questions. Self-service analytics brought easier visual interfaces that helped tackle the technical skills gap, but even with an intuitive visual approach, the ability to ask the right questions remains. AI, particularly machine learning, affords the application's embedded analytic capability the ability to automatically interrogate far larger data sets and offer suggestions that are more prescriptive than descriptive. Married to a more human approach to user interface, both visual and increasingly incorporating voice, the employee can be proactively informed within the moment of decision.
Naturally, there is a catch. Another of the barriers to traditional top-down data-driven programs has been the necessary work and investment to ensure that data is available, of trustworthy quality, and appropriately governed – subjects that only become more important as the number of users increases. Matters such as data quality and master and metadata management still need to be tackled under the auspices of a continuous enterprise-wide data governance program. Our advice is clear: a pincer movement is required – ongoing investment in data governance across the enterprise to ensure data availability, and the evolution of enterprise applications incorporating ever greater machine intelligence to surface actionable insights – within the process – for everyday decision-making.
Tom M. Pringle, Head of Technology Research, Ovum