I have posted a new technical report on SSRN: “Critsly and StudioCrit: An Artefact-Aware AI Critique Workspace and Simulation-Based Readiness Study for Design Education.”
The 13-page report was written on 6 September 2026 and posted on 12 September 2026. It documents the current implementation of Critsly and StudioCrit, the simulation work used to assess their readiness, and the limits of the evidence available so far.
Connecting critique to the work itself
Design critique is not simply the production of feedback. It requires an understanding of the artefact in progress, the designer’s intentions, the questions under consideration, and the revisions that might follow.
Critsly is designed as an artefact-aware AI critique workspace for this process. It combines:
- A visual board
- Design-intention fields
- Guided reflection
- Perspective-based critique
- Action planning
These elements create a path from the work being developed to feedback and then to a possible next action. The goal is to make critique more situated, structured, and useful within an iterative process.
Critsly as a platform, StudioCrit as a focused mode
StudioCrit is the architecture-studio research mode built on Critsly. It demonstrates how Critsly can provide a shared platform for more specialised plugins and applications.
The core platform supports the artefact-to-action critique workflow. StudioCrit extends it with capabilities for architecture education and research, including studio and class organisation, role-based access, cognitive and architectural classification, educator analytics, and exportable evidence.
This relationship is central to the work. A common platform can support critique across domains, while a specialised mode supplies the roles, concepts, analytical views, and research structures required in a particular setting.
Evidence from simulation and rehearsal
Three simulated studio scenarios generated 109 classified evidence rows. The system assigned 85 of those rows to higher-order Bloom categories.
In a separate rehearsal using 50 disposable learner accounts, the system generated 56 evidence rows, 46 of which were assigned to higher-order categories.
A later hardening rehearsal recorded 50 completed sessions and 50 successful board pulls. It also recorded 50 denials when student accounts attempted to access educator analytics, confirming the tested role boundary in those rehearsals.
These results provide evidence that the software could execute the rehearsed critique and evidence workflow under the simulated conditions.
A necessary boundary around the findings
The report does not claim that Critsly has demonstrated improvements in learning.
The scenarios and traces were simulated or synthetic. They do not measure human learning gains, learner cognition, or classroom effectiveness. The classifications should also not be treated as validated measurements of thinking.
Automated classifications remain provisional. The current report does not establish classifier accuracy or inter-rater reliability. Consequently, a higher-order label means that the system made that assignment; it does not prove that a human learner exhibited higher-order cognition.
This distinction is essential for interpreting the work responsibly. Technical readiness can support later educational research, but it cannot replace studies involving actual learners and educators.
The next stage
The implemented workflow creates a basis for controlled evaluation and further development.
I am particularly interested in collaborations involving:
- Human-participant studies in studio or critique-based learning
- Validation of cognitive and domain-specific classifications
- Comparison with educator, peer, or expert coding
- Studies of how feedback is translated into design revisions
- Adaptation of Critsly to disciplines beyond architecture
- Practical applications that require structured, artefact-aware critique
Possible collaborators may come from design education, architecture, learning sciences, human–computer interaction, learning analytics, responsible AI, or other fields where feedback and revision are central to practice.
Critsly is intended to support investigation as well as application. StudioCrit provides one concrete example of how the platform can be adapted to a specialised setting. Future partnerships can help test that approach with people, strengthen its research foundations, and explore where else it may be useful.
Read the report on SSRN: https://ssrn.com/abstract=7425258
DOI: https://doi.org/10.2139/ssrn.7425258
If you are interested in using Critsly in research or developing an application around its critique workflow, please get in touch.
Leave a Reply