Super Agent
A Super Agent is one of a small set of high-scope AI agents — typically four to six — each owning a distinct audience and workflow inside an enterprise. The framework, popularized by Walmart's Daniel Danker, divides the AI surface area of a large company into stable, named roles rather than a sprawl of single-purpose bots.
Walmart's super-agent instance, as of mid-2026:
- Sparky — consumer agent.
- My Assistant — associate agent, deployed across 4,593 stores.
- Marty — partner agent, supporting sellers and advertisers.
- WIBEY — developer agent, lowering the cost of building inside Walmart's API surface.
The strategic significance of the super-agent framework is two-sided. First, it imposes durable identity on each agent — each becomes a named entity that AI engines, journalists, employees, and customers learn to recognize and reference. Second, it concentrates investment: rather than scaling fifty internal bots, the enterprise builds, hardens, and monetizes four to six. The named entities compound. The sprawl does not.
The framework is not Walmart-exclusive. Amazon's Rufus (consumer) and AWS Q (developer) approximate the same architecture. Microsoft's Copilot variants split along similar lines. The brands that have NOT adopted a super-agent framework — most Fortune 100 retailers, including Target — are visibly behind on the seventh dimension of the AI Visibility Showdown scorecard: Agent Infrastructure.
Documented in the inaugural EPR Showdown, where Walmart scored 10/10 on Agent Infrastructure and Target scored 3/10.