Normalized claim
Driver waiting time during cement pickup: 40% decrease
reduced driver waiting times at the plant gate during cement pickup by up to 40%
UNACEM: agentic AI logistics assistant with watsonx Orchestrate reduces cement pickup wait time 40%
Use case typeWorkflow automationUpdated Mar 30, 2026
UNACEM, a Peru-based industrial group operating in cement, aggregates, concrete and power generation, created an agentic AI logistics assistant for its Lima plant to reduce truck-driver waiting time during cement order picking and preparation. The assistant is exposed through WhatsApp and web chat and uses IBM watsonx Orchestrate as the orchestration layer, watsonx.ai for LLMs and embeddings, IBM Code Engine microservices, pgvector on IBM Databases for PostgreSQL, and IBM Cloud Object Storage for grounded retrieval over manuals and operational content.
- Organization
- UNACEM
- Industry
- Manufacturing
- Location
- Peru
- Published
- March 2026
Reported outcomes
−40%
driver waiting time during cement pickupTime & speed
Strategic outcomes
Cost efficiencyReduced congestion at the plant gateScale & capacityIncreased daily load-outsCustomer experience & trustImproved ETA reliability for customersOther strategic outcomeCreated a reusable blueprint for more agents
Why do we believe this?Outcome claims, sources, and evidence checks
Why do we believe this deployment?Customer identity, provider attribution, maturity, and source checks
- Customer
- UNACEM
- Provider
- IBM
- Maturity
- Production
- Linked source
- IBM Product Blog
ai for LLMs and embeddings, IBM Code Engine microservices, pgvector on IBM Databases for PostgreSQL, and IBM Cloud Object Storage for grounded retrieval over manuals and operational content
Customer identity supportedSource describes one deploymentMaturity supported
Primary read
Use case focus
Showing 1 of 1
- 1Workflow automation
- Truck drivers could wait up to three hours during cement order picking and preparation at the Lima plant.
- UNACEM needed to reduce queue congestion and improve ETA reliability without adding friction for frontline users.
- Built a logistics assistant exposed through WhatsApp and web chat.
- Used IBM watsonx Orchestrate to plan, route, reflect and call tools for multi-step work.
- Grounded responses in UNACEM manuals, procurement documents and operational content indexed as embeddings with pgvector.
- Used IBM Code Engine microservices for embeddings, conversation storage, channel adapters and API hooks into plant systems such as order status and plant queues.
- Reduced driver waiting times during cement pickup by up to 40%.
- Eased on-site congestion.
- Increased daily load-outs.
- Improved ETA reliability for customers.
- Established an enterprise blueprint that can be reused for additional agent use cases.
Architecture
Web and WhatsApp channels connect through adapters to watsonx Orchestrate, which coordinates planning, routing, reflection and multi-agent collaboration. watsonx.ai provides LLMs and embeddings, pgvector on IBM Databases for PostgreSQL enables vector search over enterprise knowledge stored in IBM Cloud Object Storage, and IBM Code Engine microservices handle embeddings, conversation storage, channel adapters and operational API hooks. AgentOps provides monitoring, governance and lifecycle management.
Sources & evidence1
Evidence: Medium50/100Evidence strength
- Customer explicitly identified
- Deployment status explicitly supported
- Quantified outcome available
- Technical implementation details available
Type: Blog PostPublished: Mar 30, 2026Publisher: IBMEvidence: VendorConfidence: Medium
AI-generated summary. Verify important details with the linked sources before relying on this case.
Explore related AI use cases
Customer evidence
Provider evidence
AI patterns
Browse the catalog
Was this useful?
Community
Comments
No published comments yet.