
Multi-agent text-to-insight platform. A central orchestration agent coordinates 10+ specialised sub-agents, each with scoped skills. It unifies unstructured sources (SharePoint, S3, Google Drive) with 25+ structured connectors (MySQL, Oracle, MongoDB) over a knowledge-graph layer for cross-source reasoning. Data governance and access controls restrict every sub-agent to its permitted domain. Result: USD 50M+ revenue impact and 1,000+ engineering hours saved per year.
The problem
Business users needed answers that lived across documents and dozens of databases. Every cross-source question became an engineering ticket.
What I built
One orchestrator plans the question and delegates to 10+ sub-agents, each with its own scoped skills.
Unstructured stores and 25+ structured connectors sit under a knowledge-graph layer, so agents can reason across sources.
Governance is enforced per sub-agent: each one only touches the data domain it is allowed to see.
How it works
Sources: SharePoint · S3 · Drive, MySQL · Oracle, MongoDB, 25+ connectors
Knowledge layer: Ingestion & indexing, Knowledge graph, Cross-source links
Agents: Orchestration agent, 10+ scoped sub-agents, Text-to-SQL skills
Output: Business answer, Tables & insight, Cited sources
Built-in controls: Per-agent data scopes, Access control by domain, Governed connectors.
Result
USD 50M+ revenue impact and 1,000+ engineering hours saved per year.
My role: AI Solution Architect & Lead AI Engineer
0 likes · 0 views · Nov 12, 2026
Business users needed answers that lived across documents and dozens of databases. Result: USD 50M+ revenue impact and 1,000+ engineering hours saved per year.