Content Analytics

How content analytics helps manufacturers improve product safety and save lives

Share this post:

Manufacturing problems can have a serious impact on businesses. This is especially true when these problems manifest themselves as product safety issues causing injury, or even death. Whether it’s a car or a children’s toy or an advanced medical device, product safety issues don’t just result in negative publicity about your products and services, it can also lead to millions of dollars in lawsuits and liability.

Manufacturers have to leverage all their data and external data to identify issues as quickly as possible to get ahead of avoid negative press, expensive product recalls, huge penalties by industry regulators, millions in legal liability, and most importantly, they need to protect the safety of their customers. Identifying emerging issues before they escalate into full-blown product recalls helps protect reputations, customer loyalty and money.

It’s critical for manufacturers to identify reactions to new products early, especially negative feedback and to effectively address customer perceptions and concerns as soon as possible by making sense of ALL the data, structured and unstructured, that’s available to them. And yet, most manufacturers are struggling to gain deep insight into developing trends and potential issues hidden in unstructured data like customer emails, social media posts, warranty claims, surveys, complaints to regulatory agencies, blogs posts, engineering reports, quality assurance tests, or call center recordings and transcripts.

How automobile manufacturers are proactively identifying issues

The data and tools needed to do this are more accessible than ever before. In this video demonstration, I illustrate a real-world example of how an automobile manufacturer uses IBM Watson Explorer and its Natural Language Processing (NLP) capabilities to identify leading-edge indicators for a serious vehicle safety problem, way before it escalated.

Structured data often exposes the “who,” “what” and “when” of a problem. But the “how” and “why” — often the root causes — are buried in unstructured content. Here’s an example of how manufacturers can quickly and accurately reveal the “how” and “why.”

This video shows how the automobile company is using Watson Explorer to:

  • “Read” and analyze thousands of consumer complaints
  •  Identify statistically significant trends in this data
  • Find “language” that is highly correlated to this trend, which helps identify the root cause for the problem. (Watson does this without any presupposed hypothesis of what the problem could be and without bias as to probable cause.)

The video shows how an automobile manufacturer can effectively harness text analytics on vehicle safety data to diagnose recall issues through publicly available data. The video also demonstrates how Natural Language Processing models can be created by subject matter experts at companies (not just programmers or data scientists), to effectively dimensionalize abstract concepts. This allows more teams and employees to ask questions of the data that wasn’t possible before as standard text analytics and search technology couldn’t deal with the variability in natural language text.

Auto manufacturers can now isolate and pinpoint the cause of safety issues through data from the National Highway Traffic Safety Administration (NHTSA) through basic out-of-the-box analysis tools. The same concepts can be applied to other industries and issues where unstructured or text-based data is available to manufacturers.

Getting ahead of problems by mining text for indicators

Truly understanding and managing the perceived quality of your products and getting ahead of problems requires a data-driven approach to connecting and analyzing social media, governmental and internal and external data sources to mine text for indicators, sentiment and red flags. This leads to faster issue detection, problem resolution, competitive advantages and improved product design.

Want to harness the power of all the data available to your team to identify issues earlier, resolve them faster, reduce recall and PR costs and increase sales? Learn more about how Watson Explorer can help you get started. Watson Explorer is a content analytics platform that connects and analyzes your structured and unstructured content, across systems and silos and surfaces critical insights, trends and patterns. Watson Explorer combines enterprise search with cognitive capabilities to help you explore, analyze and interpret information to improve decision-making and business outcomes across your organization.

Find out  how Watson Explorer Value Calculator can help you improve product safety and save lives through content analytics

Add Comment
No Comments

Leave a Reply

Your email address will not be published.Required fields are marked *

More Content Analytics Stories

Gain Cognitive Insights and Build Scalable Cognitive Solution with the new IBM Watson Explorer

Imagine if you could ask questions with the ease of natural language and discover insights from ALL the data in your organization without worrying about different data formats. What if you could also reveal deep insights from your data without spending hours wrangling the data! How much more productive would you be if you could […]

Continue reading

Top 5 security questions CIOs should ask before choosing an enterprise search solution

Enterprise search and content analytics have become essential for any company’s success, across industries, and especially which there is high-level of customer interaction. Users must be able to find and understand the information they need, when they need it, at the scale and speed global enterprises needs, all while maintaining security and reliability. Internal and […]

Continue reading

How Honda is using Cognitive Search to drive real changes in quality assurance

Today, unstructured information represents more than 90% of the information within organizations. This IDC Case Study estimates that the digital universe will grow 40% per year over the next decade and, by 2020 it will reach an astounding 44ZB or 44 trillion gigabytes. Most of the unstructured information exists in the form of patent filings, […]

Continue reading