Combining Human + Machine Knowledge to Make Impactful IIoT Business Decisions

Qatalyst Global - Industrial IoT USA 2017

NarrativeWave is the first intelligence platform to provide data-to-decision analytics for the Industrial Internet of Things (IIoT). The company’s software combines data with domain expertise to auto-generate actionable business insights and recommendations, providing real-time decisions on critical industrial assets.

https://www.narrativewave.com/

Transcript

1. Combining Human + Machine Knowledge to Make Impactful IIoT Business Decisions ROMAIN WURTZ – CTO
3. MYTH MORE SENSORS + MORE DATA = EASIER DECISION MAKING
4. ‣ Relevant data is more difficult to find as data grows. ‣ Too much data and not enough insights. ‣ Less than 1% of stored data is analyzed and used for decision making. (1) MYTH: MORE SENSORS + MORE DATA = EASIER DECISION MAKING SOLUTION: A system that extracts relevant facts from your data and presents them to engineers. (1) Source: McKinsey “The Internet of Things: Mapping the Value Beyond the Hype”
5. MYTH MACHINE LEARNING WILL SOLVE ALL OUR PROBLEMS
6. ‣ Not everything is applicable. ‣ Machine learning algorithms are only as good as the data they learn from. ‣ Integrating more insights adds another level of complexity. ‣ Human interaction is still needed. ‣ It’s not magic! MYTH: MACHINE LEARNING WILL SOLVE ALL OUR PROBLEMS SOLUTION: A system dedicated to the use case that analyzes data in a smarter way.
7. MYTH MACHINE LEARNING WILL REPLACE OUR ENGINEERS
8. ‣ For accurate insights, deep domain expertise is highly recommended. ‣ Machine learning algorithms are a “black box”. ‣ Engineers still have to perform root-cause analysis of the findings. ‣ Internal trust within the organization needs to be established. ‣ Explaining findings is as important as knowing what happened. MYTH: MACHINE LEARNING WILL REPLACE OUR ENGINEERS SOLUTION: A system that combines engineers’ knowledge with some level of automation.
9. MYTH SHIP IT AND FORGET IT
10. ‣ IoT analytics are not “plug and play”. ‣ Analytic models need to be easily implemented. ‣ Engineers need to be able to iterate to make the system more efficient. ‣ Manageable without an IT team. MYTH: SHIP-IT AND FORGET IT SOLUTION: A system for engineers that can be easily updated by them and gets smarter over time.
11. USE CASE ENGINES AND TURBINES
12. Industry ‣ Turbines Manufacturer ‣ Large turbines focused on power generation Size of Operation ‣ Manages over 2000+ turbines (ex. Utilities) ‣ 5+ service centers dispersed globally ‣ 40,000+ alerts per year ‣ Average alert takes 16 hours to solve CUSTOMER: TOP OEM
13. Problem ‣ 50% false positives (alarms & alerts) ‣ Repetitive service center workflow process ‣ Millions of dollars wasted per year Cause ‣ Manual processes, no automation Impact ‣ 95% time reduction ‣ 25% increase in accuracy ‣ 50-75% cost reduction ‣ Improved customer satisfaction USE CASE: SUMMARY & OUTCOME
14. MACHINES ‣ Analytics on Big Data ‣ Automation ‣ Standardization FINAL TAKEAWAYS HUMANS ‣ Subject Matter Experts ‣ Understanding ‣ Explaining HUMANS + MACHINE ‣ Automated insights ‣ Complex analysis ‣ Prediction
15. AU REVOIR!