What challenges did Shorthills AI face before AI implementation?

Deliver complete, accurate, citation-backed answers across hundreds of thousands of legal documents. Meet security and governance constraints for sensitive legal data, including regulated and potentially on-premises environments.

What was the impact of the AI solution?

4X increase in comprehensiveness. 9X increase in diversity of legal arguments. 60%+ improvement in recall and precision.

Source

Shorthills AI - Trusted Legal AI with watsonx.data hybrid search

Use case typeLegal AI assistantUpdated Mar 28, 2026

Shorthills AI built a production-grade legal AI assistant capable of delivering complete, accurate, citation-backed answers at enterprise scale. The solution uses IBM watsonx.data to support secure hybrid retrieval over hundreds of thousands of legal documents.

Organization: Shorthills AI
Industry: Legal
Location: United States
Published: March 2026

Reported outcomes

diversity of legal argumentsOther quantified impact

4xcomprehensiveness+60%recall and precision improvement

Strategic outcomes

Customer experience & trustDelivered citation-backed answers instead of partial answersRisk & complianceSupported secure enterprise deployment for regulated legal dataCost efficiencyReduced search timesEcosystem & partnershipsCreated a foundation for AI agents and downstream automation

Primary read

Use case focus

Showing 2 of 2

1Legal AI assistant
2Semantic search

Deliver complete, accurate, citation-backed answers across hundreds of thousands of legal documents.
Meet security and governance constraints for sensitive legal data, including regulated and potentially on-premises environments.

Ingest and chunk legal documents, enrich them with entity extraction, and store embeddings in IBM watsonx.data.
Route queries to keyword, vector, or graph retrieval based on the question.
Use reranking and citations to improve answer completeness and trust.

Technologies

IBM watsonx.data Langflow

4X increase in comprehensiveness.
9X increase in diversity of legal arguments.
60%+ improvement in recall and precision.

Architecture

A modular hybrid search architecture: legal documents are ingested into a secure data lake, chunked and enriched with entity extraction, embeddings are stored in IBM watsonx.data, and a routing layer selects keyword, vector, or graph search with reranking and citations.

Sources & evidence1

Groundedness: 5/5Type: Customer StoryPublished: Mar 28, 2026Publisher: IBMEvidence: PrimaryConfidence: High

Primary source

AI-generated summary. Verify important details with the linked sources before relying on this case.

Explore related AI use cases

Customer evidence