Stack AI co-founder Bernardo Aceituno talks with Sophia Dew about deploying AI agents in enterprise workflows and regulated environments.
Aceituno says his company’s internal task benchmarksA benchmark is a defined test or collection of tasks used to compare a system's performance under specified conditions. reached a level he calls enterprise AGI in February 2026. He defines this narrowly around specified automation capabilitiesWorkflow automation uses software or AI to complete a connected sequence of routine steps with less manual effort.; the interview does not independently establish general intelligence.
He discusses governance of applications and credentialsAn agent permission boundary defines and enforces which resources and actions an automated agent is allowed to use., model hosting and open-weight adoptionAn open-weight model makes its trained parameter values available so others can run or adapt it under the release's licence., friction with websites designed for people, and his expectations for more proactive agents.
Watch on YouTube
