Telling a Story with Data: How AI Predicts Facility Failures | Alex Agnello
Alex Agnello, a seasoned technology leader and the founder of True Data Ops, joins host Jay Culbert for this special episode. With a background spanning the U.S. Air Force, medical technology, and defense, Alex brings a "unicorn" blend of deep engineering expertise and approachable insight to the world of facilities management. He discusses the evolution of AI from a source of fear to a vital operational tool, explaining that while AI has been around since the 1950s, modern tools like Large Language Models (LLMs) are finally making this technology accessible and "Hollywood-ified" for everyday use.
The conversation dives deep into the practical applications of predictive analytics and IOT sensors. Alex explains how high-frequency data, analyzing thousands of samples per second, can detect issues like pump cavitation or motor failure months before they occur. He emphasizes that the true value of AI lies in its ability to translate complex graphs into a narrative, telling a "story" that allows facility managers to take proactive action rather than just reacting to fires.
Takeaways:
- Start Small, Not Big: Avoid "boiling the ocean" by picking a single project with low visibility but clear success criteria to test AI implementation.
- Use the "Insights" Prompt: When using LLMs (like ChatGPT or Gemini) with your data, use the keyword "insights" to uncover patterns or issues you might have overlooked.
- Anonymize Before You Upload: If your organization has strict data rules, ensure you anonymize sensitive information before sharing it with a public large language model.
- Request Multiple Versions: When using AI to draft professional communications, ask for three different versions so you can pick and choose the best verbiage.
- Layer Your Data: Combine BMS data with external IOT sensors (like vibration or temperature) to fill in the "gaps" and create a more accurate predictive profile for your equipment.
- Time Your Reports: For the highest engagement, push automated data narratives to your team early in the morning (5:00 AM – 6:00 AM) when facility managers have quiet time to plan their day.
- Consider Space Management: Remember that buildings exist for people; use space management data to adjust maintenance schedules based on actual room capacity and usage rather than just fixed calendar intervals.
Quote of the Show:
- "If I can tell you a story, then you're gonna understand it, and you're gonna be way more engaged rather than me looking at a graph."
Links:
- LinkedIn: https://www.linkedin.com/in/alessandroagnello/
- Website: https://truedataops.com/
- Email: aagnello@truedataops.com
- Phone Number: (617) 936-9800
Creators and Guests
