An industry outsider who sees what insiders miss.

Biography

Ernest Dwane Johnson is the founder of ToastDeck Research and HaaiWorks, and the author of the B2Ai Thesis — the foundational argument that in AI-mediated commerce, recognition and selection are not the same event.

Ernest D. Johnson

ToastDeck Research

That distinction didn’t come from a lab. It came from a career spent training people to perform under pressure across industries where the margin for error is zero — commercial transportation, tax preparation, food service, real estate, and instructional design. Ernest built things, taught people to operate them, and watched systems fail at the moment of execution. When AI agents began making business selections on behalf of users, he recognized the same failure pattern: businesses that were known, but not chosen. Visible, but not reachable.

ToastDeck Research exists to close that gap. The B2Ai Thesis and the SOMAR framework — Selection, Output Mediation, and Authority Resolution — are the intellectual infrastructure behind that work. Ernest’s conviction is simple: execution under constraint is the only kind that counts, and most businesses have no idea whether AI can actually select them when it matters most.

The B2Ai Thesis.

The B2Ai Thesis — Recognition ≠ Selection — is the foundational document establishing the distinction between AI recognition and AI selection in commercial contexts. It introduces the eight-layer operating framework, the SOMAR model, and the field measurement approach behind SelectionIQ.