A new generation of analytics tools reduce the friction between human creativity, AI and automation so that they can work better, together.

Screen Shot 2020-10-29 at 6.13.26 PM.png

14,000 knowledge workers at Bayer Crop Sciences make data-informed farming decisions every day, aided by visual analytics. Bayer also has a team of data scientists who explore big data. And Bayer automates farming tasks with robotics and drones. Their IT department manages the data.

Bayer’s approach represents the future of digital business: humans, AI and automation all working together.

At most companies, the teams that make this knowledge work happen use different tools, which makes it difficult to scale analytics.

Tomorrow, humans and technology will work better together. The four disciplines that aid insight discovery will converge into one experience.

Here's why hyperconverged analytics matters and how it will work.

AI as Input, Not the Decision Making

Lies spread faster and farther than the truth. That’s why if you ask 100 people what they think about artificial intelligence, 99 will talk about job loss. This view is largely wrong. Yes, AI will help automate some repetitive, manual, predictable human tasks; technology has been doing for over 100 years. But AI will also be a vibrant job creator, at the same time.

Jeff McMillan has it right. Jeff is a chief analytics officer at Morgan Stanley. Although his job is to create insights with data science, he’s the first to point out that the human condition is the key: "a quality, empathetic, thoughtful human being, armed with tools, unequivocally produces the best outcome for our firm.”

Nailed it!

To illustrate this point in the talk below, Jeff shares the mission of a renowned AI summer seminar at Dartmouth College. Here’s its stated goal:

We propose that a 2-month, 10-person study of artificial intelligence. The study is to proceed on the basis of the conjecture that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it. An attempt will be made to find how to make machines use language, form abstractions and concepts, solve kinds of problems now reserved for humans, and improve themselves. We think that a significant advance can be made in one or more of these problems if a carefully selected group of scientists work on it together for a summer.

McMillan points out that little or no significant progress on its goal. The seminar, called the Dartmouth Workshop, happened over sixty years ago, in 1956!

The truth is, AI, machine learning, and algorithms are inputs to analytical thinking. Scaling insights require tools that aid and inform human intuition. Jeff’s talk is one of the finest to address this issue, here in its entirety:

Analytics is a Team Sport

Jeff lays the debate to rest: quality, empathetic, thoughtful human beings, working with AI tools, produce the best decisions. So why can’t humans and AI get along? Because teams are hard to create. Baseball’s perennially under-achieving Boston Red Sox of the 1970s is often called a team with “25 cabs for 25 players.” They are a sad exemplar of a group of talented individuals that failed largely due to the absence of teamwork. Michael Jordon didn’t win a single title until Phil Jackson, Bill Cartwright and Scottie Pippen helped him forge a culture of teamwork and trust.

Analytics teams have the same challenge. “Designers and data scientists frequently appear to hail from different planets” declared Bain & Company in their 2020 technology report. They do. To combat this, McMillan embeds data scientists with investment teams. He promotes the idea that data science is input to decision making, not the decision-maker. Many companies mandate bias checks on algorithms to balance its effect: for example, Google has 28 tests for ML models it uses. New job functions are emerging that hasten the delivery of analytics-fueled insights: AI operations, data engineering, and chief data officers, to name a few.

AI innovators treat analytics as a team sport. Bayer Crop Science analytics teams work closely with the farmers to understand their land, acreage, soil, and water flow; then they work with them to uncover insights.

Michelle Lacy, data strategy lead for R&D in the Plant Biotechnology Division at Bayer Crop Science, describes their teamwork with farmers: "[We uncover observations like] these are the best seeds; these are the best plant breeds that work well in this area of your farm. Planting the right seed type — the right corn line or soybean line or what have you — that will thrive in that soil type or the amount of water or nitrogen that you have. That's precision farming."

Although much is know about teamwork in sports and business, relatively little is said about analytics team building. Bayer and Morgan Stanley are good cases to study.

Hyperconverged Analytics Help AI and Human Work Better Together

Hyper-converged analytics help bring knowledge workers and AI together. Analytics teamwork in action looks like the interactive dashboard, below. In this example, our insight goal is to find the best location for a new health clinic. A healthcare location analyst drives the exploration. Under the covers, millions of potential sites are ranked by AI. The analyst explores, spelunking for insight. Flashes of insight are uncovered, spurred by human curiosity. As sections of the map are selected, new algorithmically-ranked visual clues light up. The analyst thinks and the algorithms provide clues that lead to new insights.

This is human intuition and AI working together, as a team. The algorithms, the data science, is behind the button. Hyperconverged tools make this interaction easy.

Spotfire dashboards are immersive and enable analysts to discover unexpected insights

The challenge of hyperconverged analytics is that making this connection isn’t easy. I met with the chief analytics officer of a major bank recently, and he told me that it takes them 6-9 months to get algorithms from the data science team into production. They reprogram everything. This is where model operationalization comes in.

Model operationalization facilitate frictionless, managed, secure deployment of AI models in minutes instead of months. They help knowledge workers discover, try, and adapt data science models more quickly.

Automation Analytics

Automated systems are part of the hyperconverged team too. The screen below shows data streaming off a connected car in an automated wind-tunnel simulation for Formula One racing. The system automates the evaluation of millions of different combinations of track conditions, car configurations and driver response. Sensors embedded in the car emit data about tire pressure, suspension position and brake temperature in real-time.

*Real-time wind-tunnel simulation data from embedded sensors on a Formula One car.*

Automation analytics evaluates millions of data points in real-time. Colors, graphs, track location data adjusts thousands of times a second. A race analyst set up this view to display performance thats worse than expected in red. Green is good.

At a glance, you can see that our speed is fast but our brakes are hot. Human experts use this insight to think. It’s like watching a data driven movie. Automation analytics help humans learn from the live interplay of environments conditions and automated action.

This kind of real-time insight into a car is essential for formula one racing teams. But it’s also essential for logistics and supply chain systems that consider the impact of the weather. It's essential for health care workers to understand the vital signs of their patients in real-time so they can better decide on a course of action.

Automation will transform every business, every field, and every job. To learn more, read the birth of analytics for data in motion.

Data Engineering is a Team Sport, Too

The foundation of analytics, AI and automation is data. Today’s data landscape is complex, distributed, and fast-changing; it’s also laden with legacy and disorganized. Yet data management techniques from 30 years ago still reign: centralized databases, data warehouses and ETL. Analytics leaders democratize data.

Your New Enterprise Data Fabric explores the implications of a modern data fabric that virtualizes streaming data, AI and ML model operationalization, historical data, data quality, filtering, aggregation, governance and metadata. That architecture is shown, below.