Was this forwarded to you by a friend? Sign up, and get your own copy of the news that matters sent to your inbox every week. Sign up for the On EdTech newsletter. Interested in additional analysis? Upgrade to the On EdTech+ newsletter.

Blackboard’s plan of post-bankruptcy reorganization went into effect at the end of February, four and a half months before the Building Blackboard Together users conference on July 13–15 in Dallas. That the company pulled off a real conference in that window is the first thing worth saying. In March I wrote that this event would be the first visible readout of what the new ownership structure would mean, and that a key question was whether we would see stabilization or repositioning. The answer is stabilization, executed well.

That distinction carries more weight than it sounds. The message this conference had to land was we’re back. The message it could easily have landed instead was we’re alive—a company going through the motions because not having a conference would have looked worse. Blackboard cleared that bar, and the clearest evidence was people rather than product. Attendees were engaged, several announcements landed well, and the company is visibly hiring senior leaders (I joked at one point that I was at the Parchment users conference). I previously noted that bankruptcy voided the old executive agreements and let new ownership rebuild the leadership framework from scratch. This is what that looks like.

At the same time, this is a company in a holding pattern, and there was not a ton of meat in the announcements. Matt Pittinsky—Blackboard co-founder, later Parchment CEO, and the incoming chief executive—cannot start until November due the restrictions from Instructure (owner of Parchment and a few lawyers) took longer to resolve than I expected. Our spring report put the problem this way: what Blackboard lacked heading into mid-2026 was a visible forward strategy that customers, prospects, and competitors could read, and the stretch through this conference and Pittinsky’s arrival would set the tone for the next few years. What the conference showed was a deliberate handful of features, delivered cleanly, rather than an answer to what comes next. That restraint is defensible. It also spends time the company does not obviously have.

It is worth being precise about where Blackboard sits, because the headline share number understates the problem. In competitive selections, the real contest has been Canvas versus Brightspace for years, with Blackboard’s role reduced to persuading customers to stay put—true before the bankruptcy and true today with the most notable exception with Centennial College. What changed in February means that Blackboard has its best chance in a decade to change that market situation: no debt, a single product-line focus, new capital, new ownership. They are not Canvas and they are not Brightspace. They are also a company with 1,200 institutions and 16.5 million learners that is, for the first time in years, fresh and focused on teaching & learning.

Below: the core LMS work, Ally, the AI portfolio, the Cursive acquisition, security, and the refreshing argument made from the conference stage against the curse of credentialism.

Learn Ultra course adoption is past 86% globally, and the end of Blackboard Learn Original is more or less a fait accompli (as it should have been years earlier). The remaining Original courses stop running at the end of this year, with conversion to Ultra and a read-only playback environment following. For a transition that has run the better part of a decade, and consumed an enormous amount of institutional attention along the way, the finish line is close enough to plan around.

The roadmap sessions were a long list of everyday improvements, almost every item carrying a provenance: community votes, Ideas Exchange counts, or Early Adopter feedback. This is where Blackboard continues to deliver.

The one announcement that changes what customers get rather than how it looks is outcomes. Outcome and competency alignment, longitudinal trend data, and mastery insights are now included in the core Blackboard license rather than sold alongside it.

My read: this is a company finishing a Dilbert-esque decade-long platform transition while its competitors are continuing to move forward, and there is no way around that. But the backlog-clearing does a second job. After eighteen months in which customers had every reason to wonder whether the company would still have a real development organization, shipping the specific things people voted for answers that question more credibly than a strategy slide would. The pricing move is the more interesting one, because it runs against what Instructure and D2L are both doing with new, higher pricing tiers. Blackboard went the other way, folding outcomes into the core license and shipping PDF auto-tagging at no additional charge. In a retention year that is a rational trade. It is not a posture that survives AI compute costs indefinitely, however.

Ally has been embedded in institutional workflows for years and now runs in roughly 1,000 institutions. Two numbers from the conference matter more than that count: Blackboard doubled the size of the Ally development team this year, and 60% of Ally institutions run a competitor’s LMS.

The news was PDF auto-tagging, now generally available after an early adopter program that drew more than 70 institutions against a target of 20. Untagged PDFs are the most common accessibility failure in course content, and remediating them has traditionally meant Acrobat Pro licenses and document-by-document work. Auto-tagging does it inside the Ally instructor feedback workflow at no additional cost under the existing license, with the instructor reviewing and approving the tags before the file returns to the course. This announcement was the highlight of the conference keynotes.

The rest runs in the same direction. Coming very soon: an AI alt-text assistant for images inside PDFs, and an enhanced reader alternative format with word emphasis, a reading ruler, and dark mode. Further out: bulk handling of decorative images, guidance for improving already-OCRed PDFs, and rebuilt accessibility reporting.

The accessibility work also runs outside Ally. Audio descriptions for video are available now under the core license, and the Learning Object Repository carries accessibility scoring, so a fix to a shared object flows to every course using it.

My read: this is the one part of the portfolio where Blackboard is not playing catch-up, and the timing is favorable. Our spring report argued that the April 2027 ADA Title II date turns accessibility from a documentation exercise—VPATs and attestations—into an operational product requirement, and that vendors without scalable remediation tooling face higher churn risk. The two gaps we named as largest were untagged documents and audio description for video. Blackboard has now shipped both. A tool that removes a license cost and a manual workflow at the same time is an easy internal sell, and this is the piece of the portfolio least dependent on what happens in November.

Blackboard reports 10 million interactions with its AI Virtual Assistant (AVA) to date and roughly a 50% acceptance rate on AI suggestions, a rate the company says has held for two to three years. AVA now gets a dedicated panel inside the course, deliberately separated from human messaging, plus course-grounded flashcards, on-the-fly quizzing from course documents, and a rubric check that gives students pre-submission feedback while showing the instructor the full revision trail.

The genuinely new item is Scholar, a learner-facing AI study tool launching with a select cohort this fall, built by a team drawn largely from outside EdTech. Blackboard framed it around students who struggle with executive function and who already use AI through dozens of competing study apps, then said noticeably less about the product than about anything else on the stage. The overlap with D2L’s Lumi Learner Mode is worth noting.

My read: our spring report credited Blackboard with the most developed deployment of AI tools in the market and the greatest customer adoption of them, with genuine institutional learning behind it. The question we posed for 2026 was whether it would convert that latent strength into momentum. On this evidence, not yet. The AVA additions are sensible and incremental, and they land in a market where Instructure is betting on a platform-wide agent and D2L is betting on AI as an accelerant for content at scale. Both of those are arguments about where the product is going. This was a feature list.

The one place Blackboard spent new money was academic integrity. The company acquired Cursive Technology and launched the product as Cursive by Blackboard.

The pitch opened by dismissing the three existing approaches. Plagiarism detection was built to catch copying, not generation. AI classifiers are statistical guesses whose accuracy degrades with every new model and which are biased against non-native speakers. Blue books work but do not scale, and were described from the stage as a retreat rather than a policy. Cursive instead captures a continuous record of the writing process—keystrokes, pace, revisions, time on task—from first draft to submission, producing something an instructor can review rather than a probability score. Processing runs on the student’s device, and Blackboard says what a student writes stays private during creation. The company reports use at more than 100 institutions.

Cursive is also LMS-agnostic, similar to how Blackboard handles Ally. Blackboard’s own materials list Canvas, Brightspace, Moodle, Google Docs, and Microsoft 365, though today it connects through a browser extension rather than a native integration. That gives the sales team a legitimate reason to open conversations at institutions running competitors’ platforms. A company that badly needs to defend its own base now has something to sell into everyone else’s.

My read: the framing as authorship verification rather than detection is smart positioning and probably the right pedagogical instinct. It is also going to be tested. Our spring report argued that the vendors who pull ahead over the next two years will be the ones that help institutions decide where AI belongs—formative feedback, accessibility remediation, course design, learner support—and where it does not, naming surveillance overreach specifically. Cursive sits on the line between those two lists. Continuous keystroke capture is privacy-sensitive regardless of where the processing happens, and the distance between “proof of effort” and “surveillance” is a matter of institutional policy, not product design. The on-device architecture is a real answer to part of that concern. Whether faculty senates and student governments accept it is a different question, and one that will not be settled by a conference session.

Blackboard ran a dedicated privacy and security session with the CISO, the privacy and trustworthy AI officer, associate general counsel, and the chief of staff. The content was framework and certification: ISO 27001, 27017, and 27701; FedRAMP Moderate since 2020; a DoD Impact Level 4 provisional authorization added last year; annual SOC 2 reports available under NDA; signatory to the EU AI Pact. That is all well and good if you want to be the EU of EdTech, but in the real world the question is about the platform’s security and the company’s ability to respond to what is guaranteed to be an increase in AI-enabled attacks. No one can claim full immunity, and over-reliance on badges can get in the way of genuinely rethinking platform security.

Sitting in that room a week after D2L used its own trust track to make architectural claims that a customer could evaluate—tenant isolation, egress monitoring—the contrast was hard to miss. Both companies are answering the same market question, raised by the same competitor’s breach.

The opening keynote went to David Labaree, the Stanford education historian whose books—How to Succeed in School Without Really Learning, Someone Has to Fail, and A Perfect Mess—argue that American education has substituted credentials for learning, that public schooling carries an unresolvable tension between equal opportunity and individual advantage, and that the roughly 4,700-institution sprawl of US higher education is a feature rather than a flaw.

It was wonky, and better suited to me than to a ballroom. It was also an unusual but refreshing topic for an EdTech company to put on its main stage. Vendors in this market sell into the credentialing machinery—assessment, analytics, outcomes reporting, transcripts—and platforming a sustained argument that the machinery has crowded out what it was supposed to measure is not the safe keynote choice.

The Day 2 fireside put the incoming CEO on stage with Lev Gonick (ASU’s Enterprise CIO and a Blackboard independent board member) under an explicit constraint: Pittinsky is still in a cooling-off period and did not speak to Blackboard’s forward direction. What he gave instead was the argument from his LMS at 30 series at On EdTech, written before he was recruited to the CEO role, and his follow-up working drafts. The original theory of change was to make it “obscenely easy” for instructors to put courses online and let the second-order effects follow. Thirty years on they have not, and he spread the blame around: vendors, his own company’s acquisition strategy, RFPs that rewarded deeper features inside existing surfaces, and institutions that do not volunteer for transformation. The system is still a course management system, and on his framing it has not yet earned the L.

On AI he was deliberately deflationary—an ordinary technology that can do extraordinary things, not a ghost in the machine—and treated demographics as the actually disruptive force, including the roughly 41 million US adults holding college credit and no degree. His candidate for AI’s first real application is assessment rather than content generation: address the integrity fear and the path clears for everything else, while feedback indexed to specific learning objectives tightens the link between a credential and demonstrated mastery. He called himself a radical incrementalist, with Parchment as the model.

My read: at a conference where the product story was deliberately modest, the most interesting things Blackboard said were about learning itself. Both stage arguments were too wonky for a ballroom and heavily US-centric, and I do not want to over-read a speaker booking. But it is good that the conference spent this much time with this topic framing, and on the critique of credentialism.

In March I wrote that the financial reset was complete and the strategic reset was still to come. Four months later, that remains the accurate summary. I also said to watch three things at this conference: the product roadmap, the sales tone, and whether the message was stabilization or repositioning.

The roadmap is modest and mostly backlog-clearing, with real strength in accessibility and one targeted acquisition. The tone was customer-focused rather than defensive, which is not nothing for a company that spent a year being used as a competitive talking point. And the answer to the third question is stabilization, chosen deliberately and delivered competently.

The risk is straightforward. Blackboard has five months before its actual strategy begins, in a market where Instructure and D2L are both making public arguments about where the LMS goes next. What matters in the next two years is client retention, read through the renewal cycle over the back half of this year and early next year. If it holds, a debt-free Blackboard is a credible competitor with room to invest. If it does not, well then I have some interesting future posts to write.

What I can say is that this conference did the job it needed to do. Blackboard is not Canvas and it is not Brightspace, and it is not going to be either of those things next year. But Blackboard is back. That is a real change from where this company was twelve months ago, and it earns the next year of attention.

The main On EdTech newsletter is free to share in part or in whole. All we ask is attribution.

Thanks for being a subscriber.