Novity
Not specified
Industrial Manufacturing
Amazon Bedrock, AWS Bedrock Embedding Models, AWS Lambda, Amazon API Gateway, AWS IAM, Amazon Aurora PostgreSQL (pgvector) [or Pinecone]
Novity
Not specified
Industrial Manufacturing
Amazon Bedrock, AWS Bedrock Embedding Models, AWS Lambda, Amazon API Gateway, AWS IAM, Amazon Aurora PostgreSQL (pgvector) [or Pinecone]
Novity helps industrial organizations reduce unplanned downtime by monitoring equipment health and detecting faults early. While Novity could identify faults and provide diagnoses, maintenance teams still spent significant time translating technical findings into clear actions. Avahi built an Al-powered maintenance recommendation system that interprets fault diagnoses, retrieves relevant guidance from OEM and maintenance documentation, and generates assetspecific maintenance actions with justification, likely root causes, and confidence levels. The result was faster, more consistent maintenance planning and a smoother handoff from diagnosis to execution.
Novity is an industrial analytics company serving manufacturing and oil and gas organizations. Its platform enables maintenance teams to shift from reactive repairs to predictive maintenance through equipment monitoring, early fault detection, and diagnostic insights that support higher uptime and operational reliability.
Novity’s diagnostics could surface what was wrong, but not necessarily what to do next. Maintenance technicians, operations staff, and planners needed recommendations that were specific to the asset, aligned to operational history, and grounded in proven procedures and best practices.
Without contextual guidance, teams risked slower response times, inconsistent maintenance decisions, and heavier reliance on scarce expert knowledge, all of which can extend downtime and increase cost. Novity needed to convert fault diagnoses into immediate, defensible maintenance actions that maintenance teams could trust and act on.
AWS provided the right foundation to combine generative Al with secure, scalable retrieval across technical documentation. Amazon Bedrock enabled Novity to leverage foundation models for natural language understanding and recommendation generation without managing model infrastructure.
AWS also made it practical to operationalize a retrieval-driven workflow through serverless orchestration, secure access controls, and API-based integration into existing maintenance management systems.
Novity chose Avahi for expertise building productionready, retrieval-augmented Al systems on AWS that turn complex technical inputs into clear, structured outputs for business users. Avahi understood the requirements of industrial maintenance workflows, including the need for traceable recommendations, consistent formatting, and confidence signaling.
Avahi also brought experience designing semantic search over technical sources like OEM manuals and procedures, ensuring recommendations were grounded in relevant documentation and aligned to how maintenance teams work day to day.
Avahi implemented an Al-powered maintenance recommendation workflow that starts with Novity’s diagnostic fault data and converts it into natural language context the system can reason over. The workflow then retrieves supporting guidance from a semantic knowledge base built from OEM manuals, maintenance procedures, and historical maintenance records.
To power semantic retrieval across technical documentation, Avahi used AWS Bedrock embedding models to create vector representations of source materials. These vectors were stored in a vector database, using Amazon Aurora PostgreSQL (pgvector) or Pinecone, enabling highrelevance retrieval for specific assets, components, symptoms, and failure modes.
Using Amazon Bedrock foundation models, the recommendation engine generates contextual maintenance actions that consider asset type, operational history, and industry best practices. Outputs include likely root causes to support troubleshooting, plus references to the retrieved source material so technicians can validate the recommendation and understand the rationale.
The system delivers structured results in a consistent schema (Asset ID, Diagnosis, Maintenance Actions, Root Causes, Confidence levels) to reduce ambiguity and speed execution. AWS Lambda orchestrates the end-to-end workflow, and Amazon API Gateway exposes REST endpoints for integration with Novity’s maintenance management systems. AWS IAM enforces least-privilege access and secures service interactions.
Novity improved the path from fault detection to maintenance execution by automatically translating diagnoses into clear, actionable recommendations with justification and confidence signaling. Maintenance teams gained faster, more consistent guidance that reduced the manual effort required to interpret technical faults and build maintenance plans, while supporting integration into existing maintenance workflows.
Let’s explore your high-impact AI opportunities together in a complimentary session
Let’s explore your high-impact AI opportunities together in a complimentary session
Let’s explore your high-impact AI opportunities together in a complimentary half-day session
