The record stays on legacy
Orders, accounts, inventory, and policy data still live on IBM i, Db2, Oracle, and other systems that generic AI integrations rarely understand.
Solution · Legacy databases
Your oldest systems can take part in your newest AI workflows.
Gluesync already reaches IBM i, Db2, Oracle, and other systems modern AI stacks struggle to access. AI Hub brings governed agent workflows to that same data plane in Gluesync 2.3.
AI Hub capabilities are coming in Gluesync 2.3
The reality
The systems AI overlooks often hold the answer
Copying legacy records into a warehouse or vector store may make them easier to reach, but it also creates lag, cost, and another access model to govern.
Every copy starts aging
A new warehouse extract or vector pipeline is current only at load time—and becomes another system your team must secure and reconcile.
Direct access breaks the boundary
Giving an assistant fresh database credentials sidesteps the users, roles, and pipeline scope already defined in the Gluesync control plane.
The approach
Keep the system of record. Add a governed path to it.
Use the Gluesync agents and permissions already close to the data. Query Forge is available today; the AI Hub surfaces below are planned for Gluesync 2.3.
- 01
Reach the source in place
Gluesync agents already sit beside heterogeneous systems and expose supported query paths through Core Hub—without a rip-and-replace project.
- 02
Bring a model, not a new identity
AI Studio uses a model you configure, while Core Hub MCP invokes tools with the caller’s existing permissions and an agent-specific allow-list.
- 03
Ask, inspect, then act
Query Studio AI drafts SQL for review. AI Studio shows its tool calls. SQL execution remains read-only by default instead of becoming an invisible autonomous step.
Gluesync 2.3 preview · Your infrastructure · Your permissions