How Hawaiian Electric Uses Databricks Agent Bricks to Accelerate Regulatory Document Retrieval
Hawaiian Electric Company, the utility serving 95% of Hawaii's population, deployed Databricks Agent Bricks to build an AI-powered Knowledge Assistant that searches and retrieves answers across 60,000+ regulatory documents. The system improved retrieval accuracy from 30% to 85% and delivers results 3x faster, enabling engineers and analysts to quickly surface relevant compliance and operational guidance without manually searching thousands of files.
Impact
30% to 85%
Improvement in document retrieval accuracy
3x faster
Improvement in document retrieval speed
Challenge
Hawaiian Electric's engineers and analysts spent significant time manually searching 60,000+ regulatory documents to find compliance guidance and operational standards, with existing search tools delivering only 30% retrieval accuracy — slowing decision-making for a utility serving 95% of Hawaii's population.
Solution
Databricks Agent Bricks was used to build a Knowledge Assistant that applies AI-powered retrieval to Hawaiian Electric's regulatory document corpus, with Databricks Unity Catalog providing governed access to the document index and Agent Bricks enabling natural language querying with source-attributed results.
Tools & Technologies
What Leaders Say
“We can now search through 60,000 regulatory documents and get the exact answer we need in minutes instead of hours.”
“The accuracy went from about 30% to 85%, which is a dramatic improvement in the quality of the answers we're getting.”
Sign up to read complete case studies, access detailed metrics, and unlock all use cases.
Full Story
Hawaiian Electric operates a complex regulatory environment, managing compliance documentation, operating procedures, and engineering standards across tens of thousands of files accumulated over decades of utility operations. For a company serving 95% of Hawaii's population, accurate and fast access to regulatory information isn't a convenience — it determines how quickly engineers can resolve compliance questions, make operational decisions, and respond to grid events.
The challenge was fundamental: the company's document library had grown to over 60,000 regulatory files, making manual search increasingly impractical. Engineers and analysts spending hours tracing through documents to find applicable standards or procedures was a direct drag on productivity and decision speed. Existing keyword search tools returned too much noise and too little precision.
Hawaiian Electric selected Databricks Agent Bricks to build a Knowledge Assistant that uses AI to search and retrieve answers from its regulatory document corpus. The system was built on Databricks' unified data and AI platform, with Unity Catalog providing governed access to the document index. Agent Bricks enabled the team to create an agentic retrieval workflow that understands the intent of natural language queries and surfaces the most relevant regulatory content with source attribution.
The operational results validated the approach immediately. Retrieval accuracy improved from 30% to 85% — a nearly 3x improvement in answer quality. Response times dropped to 3x faster than previous methods, compressing hours of manual search into minutes. The Knowledge Assistant is now used across the engineering and operations organization to accelerate compliance review and reduce the time engineers spend hunting for procedural guidance.
For Hawaiian Electric, the deployment represents a broader shift toward AI-assisted regulatory work at a utility where the volume and complexity of compliance documentation had outpaced manual navigation. The company is continuing to expand the system's coverage and exploring additional agentic workflows powered by Databricks.
Amazon SageMaker