IBM: worker-built AI agents will outpace governance

Aziza frames the challenge less as a tooling gap than an operational one. Giving employees the means to build agents is, in his telling, both an opportunity and a management problem, because the volume of agents will outrun central control. His answer is a "continuous system" rather than a compliance checkpoint, with alerts that let organizations respond when something goes wrong.

The sovereignty argument follows a similar logic. Aziza splits it into data, technology, operations and regulation, and warns that an incomplete answer can push a company into a silo. IBM's Sovereign Core is presented as the layer that spans those dimensions, with more than 200 compliance frameworks mapped to controls. The practical advice for buyers is to test how a partner's capabilities work together, not whether one product checks a box.

TechXchange, running Oct. 26–29 in Atlanta, is where IBM wants those questions aired by practitioners rather than vendors. Aziza points to customer examples from Citigroup, CVS Health and DoorDash, and says the majority of conference content is customer content — a format meant to help leaders compare design choices with peers, especially as models and tools keep changing.