Instructure brought roughly 3,000 attendees to Louisville from July 21 to 23 under the theme “Education in the Making.” The announcements arrived quickly: a new venture studio called Instructure Foundry, Foundry’s first announced project, national survey research on AI use in education, and a Canvas roadmap spanning more than 20 learner, educator, and administrator capability areas. Read together, they make a single argument. Instructure no longer wants to be evaluated as an LMS vendor. It wants to be evaluated as a learning ecosystem. The strategy is unfolding on two tracks: rapid experimentation through Foundry and more deliberate modernization within the core Canvas platform.
But the most consequential moment of the week had nothing to do with AI.
CEO Steve Daly addressed the April cybersecurity incident from the main stage rather than routing it to a breakout session or a support page. He committed to transparency, honest communication, and rebuilding trust through consistent action. Chief Marketing Officer Armin Molavi went further and told the room that Instructure cannot promise a world without incidents, and that future attacks are likely to be more frequent and more damaging.
That is an unusual thing for a vendor to say out loud at its own user conference, and it deserved recognition. The commitments that followed were specific: defense at every layer, annual verification, external audits, faster detection and remediation, more frequent external validation, and more open collaboration with customers. Two practical capabilities were announced alongside them, the ability to download courses and the ability to fully remove non-essential data from Canvas.
The communication failures are the part that still needs refinement. Institutions spent much of May working from status page updates while customer-specific findings were being validated, which left CIOs answering questions from presidents, general counsel, and worried parents without institution-level facts in hand. A disclosure timeline that works for a vendor’s forensics process does not work for a campus crisis communication plan, and the gap between the two is where trust eroded. Naming that in Louisville was the right call. Sustaining it is a different job, and it will be judged the next time something goes wrong, not this week.
Instructure Foundry is designed to give small, AI-native teams room to pursue opportunities adjacent to the core business without forcing every experiment into the Canvas product roadmap. Leadership was candid about the reason: the pace of change has outrun the conventional roadmap, and pushing innovation into core products faster than institutions can absorb it is not the answer.
Foundry’s first announced project is Project Athena, an AI study coach currently under development that is designed to generate personalized quizzes, study guides, and coaching prompts from the course materials, assignments, and upcoming assessments a student can already see in Canvas. It uses a mastery model rather than a completion model. Instructure cites Fordham Institute research on instructionally aligned tutoring to support the central claim, which is that study support tied to real course content outperforms generic tutoring. The privacy commitments are explicit: read-only access governed by the student’s existing permissions, no use of student conversations to train external models, student ownership of their data, and Instructure acting as a FERPA-designated school official.
Athena is therefore more than a study tool. It is an early test of whether Instructure can use the data and context already inside Canvas to deliver differentiated AI experiences without turning the LMS itself into an uncontrolled experimentation layer.
The Canvas roadmap is sequenced across 2026 and 2027. Nearer-term priorities include the learner dashboard, AI-assisted content creation, submission and feedback workflows, educator dashboards, communication, search, and administrative analytics. Outcomes tracking and an integrated portfolio extend the roadmap into 2027. Notebook and the new Learner Dashboard are available now across all tiers, with portfolio enhancements generally available in September.
Watch what sits behind which tier. The Block Content Editor enters an early adopter program in September with block authoring for all tiers, but its AI-powered efficiencies are reserved for Canvas Plus and Canvas Next. The roadmap story and the licensing story are not the same story.
The survey supplies the counterweight. Of 1,125 respondents, 90 percent of higher education students report using AI in class at least occasionally, compared with 61 percent of instructors. Yet 41 percent of higher education educators report no formal AI training, only 11 percent report comprehensive training, and average confidence sits at 2.60 on a five-point scale.
Three items belong on the near-term institutional agenda: a tier entitlement map, a defined approval path for Foundry products, and a funded faculty AI training plan. The first two protect the budget. The third determines whether any of this reaches teaching.
Instructure showed that it has no shortage of ambition. Whether that ambition produces institutional value will depend less on the volume of AI announcements than on trust, licensing clarity, and the readiness of educators to use what is being built.
Be sure to follow Matt Winn on LinkedIn to catch all his great industry insights.
Share Article:
© Copyright 2026, The Tambellini Group. All Rights Reserved.
Get exclusive access to higher education analysts, rich research, premium publications, and advisory services.
