| INDUSTRY
FIELD SERVICES SOFTWARE |
AI SERVICE
LARGE LANGUAGE MODELS |
Recommending Information to Field Technicians Using LLMs
Learn how Synaptiq supported a field services software company by using machine learning to make suggestions to streamline technicians’ issue diagnosis workflows.
Problem:
HVAC technicians get called out to service equipment but spend 30-60 minutes on-site searching through their field service applications for content needed to diagnose and fix particular problems before they even begin their repairs. They have hundreds of thousands of unorganized technical manuals, which are often 100+ pages.
This field services software company asked for Synaptiq’s help to digest field technician notes and use large language models to suggest and diagnose issues with the goal of streamlining technician user experiences.
How can AI in field services software spark ideas for your business?
Identifying user problems and solving them iteratively with data, machine learning, and AI is not a use case unique to field services software. And the demand for applying machine learning technologies to real-world problems is higher than ever.
Yet, getting started and understanding the nuances of how to implement AI tools is no easy task. That’s where expert consultants like Synaptiq can make a world of a difference for your business, and work with you as we worked with this field services software company to explore possible strategies and design a user-centered solution.
Solution:
Synaptiq used an LLM to extract key information like make, model, cause of failure, etc from millions of historic technician notes and created a structured database of this information.
We also used an LLM to summarize relevant troubleshooting steps and recommend them to technicians, rather than forcing them to search through long technical manuals.
Outcome:
Based on our POC Design, a technician simply scans a machine’s dataplate (containing it’s model # and serial #) and is immediately presented with a list of the most common root causes of failure for that particular model and the most relevant snippets of technical documents to help them quickly fix the machine, saving them hours per day and enabling more service runs each day.
HUMANKIND OF IMPACT
AI IS HOW WE DO IT,