What Does Augmented Analytics Mean for Procurement?
Posted by Amenallah Reghimi, Vice President of Product Management, JAGGAER in Source-to-Pay on March 24, 2020
It’s OK, we understand. Here we are throwing another buzzword at you. But bear with us because it is not really the words that matter, but the thinking and the technology behind the words. And most importantly, what augmented analytics can do for you.
Let’s step back first.
Analytics has been with us for some time – more than a couple of decades – though until recently it was generally called business intelligence (BI). But it is worth reminding ourselves of what it means, and what made it possible. Analytics is a very broad term and we can get a better understanding of what it means by viewing it in four layers, each of which answers a successively more difficult question:
- What happened? (Descriptive Analytics, formerly called Business Intelligence)
- Why did it happen? (Diagnostic Analytics)
- What might happen? (Predictive Analytics, also referred to as Augmented Analytics)
- What should I do about it? (Prescriptive Analytics, also referred to as Augmented Analytics)
Augmented Analytics was made possible thanks to the use of machine learning techniques such as neural networks and multiple regression analysis and mathematical approaches such as decision tree analysis Bayesian inference modeling. These found their earliest commercial applications in areas such as customer profiling, market segmentation and credit scoring as well as risk analysis (e.g. to determine the likelihood that a major client will default on a bank loan).
This was all very well but unless you had a PhD in statistics or mathematics or an applied mathematics subject like risk management, analytics was not a whole lot of use to you; even if you could make sense of the complex algorithms, it was hard to work out which tools to apply. And even if you got that far, you would also need to understand the underlying data and how to extract the subsets you needed from the data source.
And then you would still need to sort out which insights to act on, because, even with modern business intelligence platforms to front-end the analytics engines, insights are not contextualized or easily consumable.
Further technological developments have changed that to give us what Gartner defines as augmented analytics, which “… uses machine learning and artificial intelligence techniques to transform how analytics content is developed, consumed and shared. Data and analytics leaders should plan to adopt augmented analytics as platform capabilities mature.”
Gartner continues: “Augmented analytics is transforming how business people explore, analyze and act on insights from A&BI [Analytics & Business Intelligence] and DSML [Data Science and Machine Learning] platforms through machine learning (ML) and artificial intelligence (AI)-assisted data preparation, insight generation, model selection and insight explanation.”
In plain English, this means that analytics will increasingly cease to be the preserve of the data scientists and stats PhDs and will be more accessible to business people looking for solutions to their day-to-day and strategic challenges. And that, of course, includes procurement professionals.
Gartner also comes with a prediction: “By 2021, augmented analytics will be a dominant driver of new purchases of A&BI and DSML platforms, and of embedded analytics.”
So, let’s examine Gartner’s words in more detail, but from the specific perspective of procurement.
Machine Learning (ML) and Artificial Intelligence (AI)-Assisted Data Preparation.
As is well known, procurement generates massive volumes of data. Traditionally, if you wanted to get at this to perform advanced analytics you would first need someone (typically a data scientist or programmer) to write a special extract-transform-load [ETL] routine to get the data you need from various sources, put it in the right format, and load it into a data warehouse or data mart.
Today, in contrast to the rigid file formats of RDBMS, technologies are based on the concept of a data lake, whose content is fluid, with no boundaries and therefore no technical limits to its expansion. On top of this logical (as opposed to physical) data models make data preparation much easier. All forms of data can be pooled into one single repository where business users can interact with it in multiple ways for analytical purposes.
Insight Generation, Model Selection and Insight Explanation.
In the past, if a procurement professional wanted to run a query against data to determine what might happen in future (e.g. to minimize the risk of a breakdown in logistics supply chains) they would probably need to go and explain their business to a data scientist, who would go away and have a think about it and come back with a suggested model.
Augmented analytics, by contrast, is built on the principle of self-service by business users. The intelligence is already pre-programmed (embedded) and continuously learns as the data changes. Augmented analytics solutions are built out of the knowledge of domain experts (for example, JAGGAER customers and software development teams) who understand the business (and regulatory etc.) frameworks and who anticipate future changes – in order to get away from heavy reliance on IT.
AI-assisted smart assistants now enable the user to go beyond simple click-through self-service to guide you through model selection, insight explanation etc. using plain language (like Apple’s Siri or Amazon Alexa).
This is a journey, and we are still only at the early stages. Nevertheless, by embedding intelligence into all the key applications from source to pay, algorithms can set certain actions in motion based on the collected data, which trigger recommendations for the user in the system.
JAGGAER has already implemented this kind of solution in its own recommendation engine, which presents users with a range of possible options in a given scenario, with in-built feedback loops enabling the analytics to continuously improve based not just on new data but on the recommendations that users actually decided to follow.
An obstacle to more rapid progress is the lack of relevant skills in many procurement organizations. While augmented analytics reduces the need to know how the technology works, what is missing is access to the technology and an appreciation of why it should be used, i.e. the benefits it can bring, such as identifying supply market trends ahead of the competition.
Gartner estimates that on average, only 35% of people in organizations have access to analytics and BI tools. “Despite the compelling benefits of augmented analytics, which can accelerate adoption well beyond this limited level, efforts to incorporate augmented analytics encounter resistance,” writes Gartner, and this is the case for a number of reasons, including the lack of data literacy in the broader user population, distrust, concerns about job security etc. The outlook is slightly rosier in procurement: in our Digital Procurement Survey 50% of respondents rated their awareness of digitalization as “up to date” or “excellent”. But there is some way to go.
That said, we believe the “How Augmented Analytics Will Transform Your Organization: A Gartner Trend Insight Report Gartner Trend Insight Report” gives some excellent advice on gearing up for the next wave of augmented analytics, concluding that “Early adopters of augmented analytics have the potential to realize more strategic and differentiating business benefits from their analytics investments than those who wait until these technologies are widely adopted.” That is certainly already proving to be the case in procurement. The is certainly worth your consideration.
Gartner, How Augmented Analytics Will Transform Your Organization: A Gartner Trend Insight Report, Rita Sallam, 31 October 2019