augmented analytics, analytics, manufacturing, alexa, KPIs

Mozart, Metallica and Manufacturing Operations Management

Every day when Jim comes home from his work as a production planner at the factory, he ritually grabs a can of beer from the fridge, plops down on his coach and gently gives the order,  “Alexa play relaxing music.” He sips his beer, closes his eyes and within seconds one of his favorite rock ballads flows out of the speakers.

Does Jim realize that the amazing Artificial Intelligence (AI) technology underpinning Alexa is about to fundamentally change his work environment? Does he know that soon on the shop floor he will be able to gently ask questions and receive answers from an Alexa-like chatbot within seconds? And does he know that he is entering the era of Augmented Analytics?

What is Augmented Analytics?

Augmented Analytics is an approach of data analytics that employs the use of Machine Learning (ML) technology and natural language processing (NLP) to automate analysis processes normally done by a specialist or data scientist. It uses smart algorithms that scan data and detect patterns, trends, exceptions and correlations to provide meaningful insights, reporting, and notifications.

Will Augmented Analytics Replace the Analytics Solutions that We are Familiar with Today?

What comes into most peoples’ minds when thinking about analytics, and my mind as well, are dashboards. The modern dashboard includes interactive charts and graphs that visualize Key Performance Indicators (KPIs) and offer the ability to drill down into related data.

Until recently, in this twilight of the Visual Data Discovery era of analytics, that was indeed the most successful form of analytics. Given, however, the increasing volume of data with greater variety (“Big Data”), we are reaching the limits of what us humans can effectively discover visually. It is impossible to visually monitor thousands of metrics and, because data used in analytics is increasingly real-time, those thousands of metrics change often. Thus, slowly but surely we are reaching the end of the Visual Data Discovery era.

Now and going forward, we can use software that does the discovery job for us; it automates the detection of trends and exceptions. By using artificial intelligence and smart algorithms, Augmented Analytics can proactively alert you when unexpected or potentially interesting things happen. 

What is unexpected or potentially interesting? That is something the algorithms must learn, which is why they use Machine Learning (ML) technology; with each iteration the algorithm becomes better at doing its job. Therefore, it needs to obtain feedback about the success or failure of the outcome, including feedback from you as a user. This is just like Alexa learning about what your favorite music is so that when YOU say “relaxing music”, Alexa knows to play a slow Metallica song instead of Mozart. Augmented Analytics also means that you will interact with the system by asking questions and getting answers in spoken or written language. Augmented Analytics uses “natural language processing” (NLP) which includes “natural language query” (NLQ) to understand your questions and learns to find answers, and “natural language generation” (NLG) that provides answers in spoken language and/or produces alerts.

In summary, we are entering the era of Augmented Analytics, which uses AI and ML to enhance the discovery process and NLP to interact with the user.

Is Augmented Analytics the Death of Dashboards? 

Will this be the death of dashboards as we know them? The answer is probably “yes” over the longer term. Eventually, the ML-supported “decision intelligence” will overcome traditional dashboards based on “business intelligence”. Augmented Analytics will still use charts and graphs to illustrate answers, but users will no longer need to stare at dashboards for hours trying to find insights. Classic dashboards, however, will still be useful in specific situations where human interaction proves challenging. For example, dashboards on large displays on the shop floor using andon boards that indicate production status makes it easy for everyone on the shop floor to keep an eye on what is going on. These environments are not conducive to a quiet conversation with Alexa.

The Adoption of Augmented Analytics in Manufacturing

Augmented Analytics is already broadly used in finance and market research. The analysis of consumer preferences and social media trends helps decision makers make B2C positioning and targeting decisions for their markets. Tracking consumer activity on the web results in large volumes of fuzzy data and Augmented Analytics are required to make sense of that data.

Augmented Analytics has not yet fully made a breakthrough in manufacturing, but the breakthrough is likely to happen soon. With the rise of the Industrial Internet of Things (IIoT) and integrated supply chains, manufacturers can obtain a significant competitive advantage if they succeed in gaining insights from the increasingly rich set of available data. Augmented Analytics makes it possible to not only analyse the internal data of the company but also external data from many sources along the value chain. In fact, without Augmented Analytics it is virtually impossible to reap insights from all that data. For a manufacturing company, being able to predict supply chain disruptions and quickly adapt to those disruptions, can make a world of difference. 

Another manufacturing use case is the optimization of inventory levels: When you have dozens of warehouses across the world it is challenging to produce an overview with visual data discovery only. Augmented Analytics can automatically detect inventory inefficiencies and proactively inform the relevant inventory managers. Also, with IIoT data, Augmented Analytics can detect hidden inefficiencies in production processes. These are just a few examples of use cases for Augmented Analytics in manufacturing; more will be discovered over time.

A Roadmap for Augmented Analytics

Today, most of us are still living in the era of Visual Data Discovery because that is what most analytics solutions have been focused on for many years. How will the transition from Visual Discovery to Augmented Analytics happen?

Existing solutions will be augmented step by step, applying the described technologies NLQ, NLG, ML and AI in an increasing number of scenarios.

For example, your current analytics solution already has a capability to raise an alert when a given KPI passes a threshold. Those thresholds are based on rules that users can enter.
Now imagine that next to those rules-based alerts, the system also starts issuing alerts for all events where KPIs show patterns that jump out from the ordinary and that it presents those alerts in the relevant user-context, for example, showing customer related KPIs when you work in a customer facing function. Depending on the interaction with the user (show me more / like it / don’t like it) the system will then learn what is relevant for each user or user role and fine tune future alerts.

Another example is the introduction of NLQ. You can ask the system any question in spoken language: “How many of my sales orders are overdue?” or “What is the available quantity of item P12345?” and receive the answer immediately in plain text and numbers, without having to open screens and running reports or dashboards. And next to the answer, with a single click, you can then open a chart or graph that shows the details behind that answer. Also in this case,  the system will learn based on your feedback and provide better and better answers over time, eventually beating your best subject matter experts. As analytics become smarter, you can also ask questions about the future using prediction algorithms; making use of historical facts and current trends, accurate predictions are possible.

In hindsight, the era of Visual Discovery enabled the end-user with embedded self-service analytics and was taking away the need of a data analyst to spot visual insights in the data. 

Going forward, Augmented Analytics enables the end-user even more, providing insights that previously only a data scientist could provide.

With Augmented Analytics we are on an exciting journey. It is difficult to predict how fast this will go and where it will eventually lead to, but it is sure that Augmented Analytics, combined with other technological innovations like process automation, will revolutionize the way we run businesses in the near future.

LEAVE A REPLY