Listen to our blog:

According to a recent survey, 47% of companies today consider AI/ML as a top priority in 2024. Yet, according to Harvard Business Review “cross-industry studies show, on average, less than half of an organization's structured data is actively used in making decisions—and less than 1% of its unstructured data is analyzed or even used at all.” ¹

This statistic should concern R&D or Engineering decision-makers. Companies are drowning in data, and very few companies can leverage their data because they are stuck trying to find, access, and connect various data sources. So first ask yourself these three questions:

• How can a company move forward with AI or ML if it does not have the right data?
• How does a company get its arms around today's data challenges?
• How does a company use data to foster innovation, and collaboration to gain a competitive edge and advantage in their market?

Over the past eight years of delivering data solutions to customers at Viviota, we have observed a number of data practices common among the world's most innovative and successful product companies when it comes to getting value from their data. Out of this work, we compiled the top five top actions companies should take right now to improve their operations around data and AI/ML:

1. Create a data strategy
2. Understand the value of your data
3. Understand how data will drive decisions
4. Make thoughtful technology decisions in support of strategy and decision-making
5. Understand the role of artificial intelligence (AI) and machine learning (ML)

1—Create a data strategy

Just like every company needs an overall business strategy to succeed, a data strategy is equally as critical. Before we discuss what technologies should be involved, there are many things to consider with any data strategy.

A few years ago, we had the great pleasure of engaging with a jet engine manufacturer to study engineering data. Making a jet engine is a fascinating process. First, it has over 23,000 parts. It must operate in extreme conditions. A jet engine lasts 30 to 40 years and is refurbished every five to seven years. The common thread across the engine lifecycle is data. The team at this manufacturer wanted to prove that by tapping into existing engineering data, they could reduce overall R&D costs. The ensuing study confirmed this theory.

Some initial considerations we learned in this study are good candidates to incorporate into your data strategy, including:

• Make it someone's job at the executive level to have data responsibility full-time.
• Identify the top data assets in your company and assign a value.
• Determine your data governance strategy.
• Establish high-value business impact targets that demonstrate ROI.
• Identify technology decisions concerning data that support the company's business strategy.

2— The value of data

Genuinely great companies have developed methodologies to identify high-value data. As the diagram below illustrates, when data is created, it has little to no value. As the data is incorporated into an application and begins to create insights, the value increases. The highest levels of automation can drive immense value for corporations. This is true in designing, building, marketing, and selling a product. Data is required every step of the way.

Information extraction begins in the middle layers of the Value Chain of Data model. It is here that the high-value transformation to deep insights begins. At this layer, the data has been stored in a way that can be easily searched and manipulated. In the end, understanding the journey removes obstacles and speeds the transformation to drive meaningful action.

The best companies have figured out the value of their data and put their precious resources to work mining that data and turning it into automated outcomes.

Unfortunately, companies are really good at creating data, but as the Harvard study points out, they only use about 1% of the data!! The value chain blueprint is a great way to start your data journey and build in the people, processes, and tools necessary to get the most value from your data

3—Understand how data will drive decisions and how to use AI and ML in that process

The science of making good decisions has advanced dramatically over the years. More than ever, data is being used in the decision-making process. Some decisions happen in the blink of an eye. Others evolve slowly, and time is needed to get the proper context of the situation. External factors can be an immediate or long-term driver of decisions. Psychologists have written books on the human brain's decision-making process. One tenet stands out: no matter how much data you have, there is always a human side to every decision. Over time, AI and ML are intended to make better decisions than humans. The magnitude, scope, and speed all feed into a decision funnel.

Small decisions at the base of the funnel are made quickly and easily. As the decisions get more complicated, they require management input. At the very tip of the funnel, the most challenging choices happen.

Too often, slow decision-making is a consequence of culture instead of a strategy. Companies usually have a decision-making culture that fits with where they are on a maturity and culture spectrum. Decision-making needs to be decoupled from that spectrum to some extent. Strive for a balance during the decision-making process. Good habits for this process should include:

• Ensure timely decision-making includes both the correct data as well as human input.
• Recognize that all decisions have consequences, and each company must strike the right risk/reward balance between speed and taking action.
• Form a conscious and well-communicated strategy for decision-making within the company. Build a culture and framework to support and implement the plan.

4—Make thoughtful technology decisions that support your data strategy

Today, many companies have chosen technologies to form a data repository or data lake. IT organizations must overcome the challenge of balancing people, process, technology, and budget.  As a result, the size and scope of data projects vary significantly by company. One of the more complex balancing acts has been addressing data in the Operational Technology (OT) and Information Technology (IT) domains. Mainstream cloud-based database technologies offer little in the way of specific help tailored to consumers of engineering/scientific data.

Industry-leading companies have embraced the idea that data is an asset. As a result, they have formed digital transformation teams. These teams closely examine the types of data required, the necessary velocity to process and store the data, and the transformation needed to make the data useful. With sensors and video becoming ubiquitous and technologies like LIDAR and radar being deployed, existing data repositories may have gaps. Companies, especially those concerned with OT data, need to:

• Plan with care and foresight where data is stored and processed.
• Plan carefully where raw data is stored vs. where metadata is stored.
• Think through the security and timing implications of cloud vs. edge technology
  

5—Understand the role of Artificial Intelligence and Machine Learning

Progress in AI and ML is rapidly accelerating. Many companies are building entire software engineering teams to focus on specific applications of AI and ML technologies. Numerous open-source tools and mathematical models can be applied to jumpstart efforts and solve fundamental problems addressed by AI and ML.

In our experience at Viviota, one of the biggest obstacles to AI and ML success has nothing to do with technology. Addressing social engineering, changing management practices, and being willing to make the appropriate investments are prime concerns. Considerations for this step include:

• Build vs. Buy—with human software resources at a premium, understanding what should be done in-house vs. outsourced is imperative. Ask if this information knowledge is critical to company strategy and competitive advantage.
• A top priority should be reviewing data requirements for any AI and ML project. Treat data as a high-value asset. As part of this exercise, ensure that R&D analog data is also treated as a first-class citizen. In many cases, this data is the company's core IP.
• Don't overlook Change Management—Resistance to change and technology adoption is a serious issue; creating the right environment to support the AI and ML strategy is critical to success.

Summary

Getting your data house in order can be a daunting task. This article provides a framework to spur thought, and ideas, and a method to act on your data and AI/ML initiatives. All companies are different. We counsel our customers not to try and hit a home run as they start the journey. One of the top reasons projects fail is that teams are looking for the Big Bang or the Home Run instead of taking smaller incremental steps and building momentum. It is truly a journey worth taking.

If you're on your data journey and using AI or ML or or just beginning to explore how AI and ML can be leveraged in your organization, I am happy to discuss your data challenges with you and how they might be solved; you can reach me at Barry.Hutt @Viviota.com.

¹ "What's your Data Strategy?", HBR 2017