PM Substrate
Research into persistent context and observable execution across software and AI agents. The graduate paper and poster are being reworked.
Research direction
PM Substrate explores how software and AI agents can retain useful context while making their actions and results observable across connected systems.
The fundamentals
The project brings together interface discovery, retained observations, action history, and replay. Its current direction focuses on observing actual execution and preserving durable history, rather than approving or blocking the monitored application’s actions.
Paper and poster in revision
The PM Substrate paper and poster presented for Emmanuel Akinwale’s graduate program are being reworked. Revised academic materials will be added when they are ready to be shared; no revised paper or poster file is published with this project entry.
Explore a collaboration
Interested in this research or a related application? Start a conversation with JOATLABS.
Discuss this research →