Erratum: In last week’s post, I updated the web version by removing an erroneous description of UPCEA having the most conference-heavy revenue for the three groups mentioned below. OLC has a higher reliance on conference revenue than does UPCEA.

At this year’s D2L Fusion users conference, I saw many Canadians finally appreciating the full heat of Phoenix (by the way, we’re having a real heatwave right now that is five degrees hotter). But I also saw and heard a company with a strong strategy, but too often that company creates its own barriers.

For the first time since the late 2010s, D2L has a coherent product strategy rather than a collection of features. The last time was the move to the cloud and the ground-up UI redesign that introduced the modern Brightspace experience—a clear thesis about where the product needed to go, with sustained investment behind it. Since then, D2L has added a lot of stuff: AI features through Lumi, the Creator+ authoring toolset, the H5P and Course Merchant acquisitions. Individually useful, but with no obvious thread tying them together.

While we were starting to see coherence last year, this year the thread was unmistakable. D2L now describes its product strategy in four pillars—generative AI transforming teaching and learning, high-quality content creation (native tools along with Creator+, H5P, and the new CreateSpace), product-market fit for employee and compliance training, and continuous reinvestment in core Brightspace—with trust positioned as the foundation underneath. The real strategy is in how the first two pillars combine: D2L is going after scaling engaging course content as its sweet spot, with AI as the accelerant rather than the headline. The implied customer is one that skews towards coordinated course design and top-down decision-making—institutions and systems that deliberately are looking to scale course redesigns and impact—more than laissez-faire institutions. The development investment lines up with the strategy, and my conclusion coming out of conference season is that D2L currently has the most strategic product and services plan of the major LMS vendors, period.

But this is still D2L, and the company has a chip on its shoulder that shapes how customers perceive it. That posture too often makes the company defensive, and gets in the way of scaling beyond the small but healthy growth it has today.

One example from the conference is D2L’s repeated claims of greater-than-50% win rates that extend beyond what any independent market coverage supports and without backup data. And D2L’s aggressive approach to the Canvas cybersecurity incident and competitor user conferences—direct contacts to institutions during these events, described to me in multiple conversations—left more of a bad taste in potential customers’ mouths than intended.

Both examples reflect the same instinct: managing the perception rather than letting the strategy and the results speak for themselves. And that’s a shame, because this year the strategy is worth talking about. Below, I describe the major initiatives—core Brightspace investments, the Lumi expansion including Learner Mode, CreateSpace, the ecosystem play, and the trust positioning—and what each means for D2L and the LMS market.

The description of continuous-reinvestment showed up with several core product improvements this year, and notably with release dates on the slides—a response to feedback (including ours) about vague timelines at last year’s conferences.

The Content Experience—D2L is no longer calling it “new,” with justification, as roughly 70% of customers now use it—gained an immersive full-width viewer this month, with a collapsing table of contents in September and sequential-learning visibility (locked future modules visible to learners) by year end. Outcome alignment moved from a multi-click, multi-upload process to a single curriculum screen across an entire module set—meaningful for competency-based and accreditation-driven programs.

The rest of the core investment reads as an everyday-friction list, and I mean that as a compliment. In many regards, the more the LMS fades to the background, the better for its customers.

None of this is flashy, and that’s the point. This is the pillar that answers whether a vendor is still investing in the product its customers use every day, and D2L’s answer was more convincing than most.

Before the announcement, D2L shared usage evidence for the existing Lumi portfolio of AI tools, with hundreds of thousands of educator-created items, and 98% of sessions citing course material. Lumi Insights also extended from visualizing quiz data to suggesting interventions—edit the problematic question, revise content with AI assistance, or generate personalized emails to the specific students who struggled, in one click.

The commercial context, which we covered in our (premium only) spring report, comes from D2L’s own earnings disclosures: Lumi annual recurring revenue (ARR) grew from roughly $2 million to more than $3.5 million over the final quarter of the fiscal year ending in January—against total company ARR of $219.8 million. The attach rate is above 40% among new higher ed customers, but CEO John Baker has been explicit with investors that AI functionality is not yet showing up consistently in RFPs. Lumi is real and growing, and it is also still under 2% of the business. Lumi Grow as a paid tier is the kind of premium AI packaging we argued this pricing-compressed market would force.

D2L is clearly positioning AI as a method to scale course creation and redesign, describing a 50% reduction in time to build a course (this is not a verifiable claim, but the positioning is important).

The marquee announcement was Lumi Learner Mode (beta this fall), a dedicated study space inside each course that belongs to the student. Learners can highlight content and ask for explanations or alternative examples, save responses into a personal notebook with their own annotations, and launch self-directed knowledge checks—including Socratic open-ended questions where Lumi nudges thin answers toward specificity. Lumi Tutor, previously a floating button, gets a permanent home here, as shown below.

Two design choices matter more than the feature list. Knowledge check results are visible only to the student—instructors see aggregate engagement data, not individual interactions—a deliberate choice to preserve student ownership. And responses are anchored strictly to course material with citations, refusing questions outside it, an approach D2L says it has tuned for two years on the instructor-facing tools.

The pedagogical framing is the most direct answer I have heard from an LMS vendor to the existential question of whether LMS-based AI tools matter when students have ChatGPT. D2L’s argument, supported by its Digital Promise research partnership, is that AI with structure and boundaries accelerates learning while unstructured AI use degrades it—so D2L’s answer to counterproductive tools is a structured one students actually want to use. It also sharpens a strategic contrast in the market: where Instructure is betting on a centralized agent that reasons across the whole Canvas platform, D2L appears to be more focused on embedding AI into discrete teaching and learning workflows—two different architectural bets.

But the room’s reaction was noticeably muted—no visible applause for the announcement D2L billed as something never seen in an LMS. Part of that is likely from the presentation: the introduction didn’t explain enough for customers to know how to think about the new tool. In one conversation I had at Fusion, a customer asked whether Learner Mode was competing with Course Hero or something different—a question that D2L will have to address. Even if the company gets the design right, there is still a risk if institutions buy Lumi partly on the student-facing narrative and students don’t adopt it—a risk D2L acknowledged when I raised it. And that risk is not hypothetical: the dominant campus pattern for LMS AI remains hedging and pilots, with vendor keynotes and roadmaps running well ahead of deeper institutional adoption.

CreateSpace was announced at last year’s Fusion, but there were a lot more details presented this year behind the tool: a full content management system inside Brightspace—in D2L’s words, “not a repository”—where organizations create, version, collaborate on, and synchronize reusable learning content across hundreds or thousands of courses. The tagline is “create once, improve everywhere”: edit the source version, then sync the change to every course where the content lives, with control over where updates apply and support for master/gold course workflows.

To understand where CreateSpace fits, it helps to see how D2L’s content tools now layer (my understanding based on what I heard):

Services are a key part of this content strategy. D2L’s Learning Services team has used AI for two years to modernize clients’ stale course content at scale—including a “SCORM scraper” that automatically uplifts legacy SCORM courses into rich HTML, work that used to take weeks. Note that D2L makes a related but different claim here than the one above: not build time, but internal course development costs cut by more than half, along with measurable completion gains from modernized courses.

For those of us with long memories, this is a familiar idea. Australia’s HarvestRoad Hive and the federated learning content management system (LCMS) concepts of the 2000s promised much the same thing: centralized learning content repositories with reuse and governance across courses. Those ideas were ahead of their time and did not in the end become established systems, and to a degree the market fragmented into LMS-native tools, authoring systems, and in the case of large-scale online programs, often custom development.

CreateSpace somewhat revives the LCMS idea but makes it part of the LMS toolset. On one hand, this simplifies the EdTech ecosystem, but on the other hand it would create more dependence on a single vendor.

CreateSpace is part of core Brightspace at no additional cost—but core does not mean universal. It enters opt-in early access in August and further rolls out through 2026. And there are significant pieces still being built between now and full availability. D2L reports that the closed beta was the largest in company history. And the embargoed news was a co-development partnership with University of Maryland Global Campus (UMGC), whose skills-forward use case—breaking full courses into atomic, skill-tagged “skill packages” that can be searched and reassembled when a new industry partner signs—ships with the product.

My read: CreateSpace is the fullest expression of the four-pillar strategy but not a required element of it. On paper there is a lot of possibility for improving course design and updates at scale, but there are risks involved. The idea had a lot of value in 2005, and the addition of GenAI to multiply the volume of institutionally-owned content adds even more value. And for what it’s worth, CreateSpace and the UMGC video received the most reaction during the Solutions Spotlight.

The quieter strategic release was Brightspace Apps, an app-store experience for LTI integrations rolling out this summer. Partners publish integrations; admins see the requested data permissions explained in plain language, install in a few clicks, and control deployment scope—institution-wide or a specific department. LTI 1.3 only and existing integrations unaffected. The companion Brightspace Builders program gives partners and institutional developers the documentation and publishing path with much better information presented to institutions when browsing third-party apps.

It is worth noting that Brightspace Apps acts somewhat as a counterweight to the CreateSpace approach. D2L is simultaneously making it easier to connect third-party tools and more attractive not to need them for content.

D2Link, D2L's integration product connecting Brightspace to enterprise systems like Workday and Oracle, grew from 30 to 70 clients year over year, with the company reporting significantly-reduced implementation timelines. Very few institutions are switching LMS vendors right now, so lowering the cost and drama of migration is one of the few levers available to unlock the selections that do happen.

The conference’s framing theme was “innovation with purpose, built on a foundation of trust,” with a dedicated trust track running throughout Fusion, including sessions with the CTO and SVP of Engineering. D2L was candid at the executive briefing that the trust track was re-framed and emphasized in response to the Canvas cybersecurity incident.

D2L’s engineering analysis of the breach—a thesis built from public disclosures, not inside knowledge—is that profit-motivated attackers were, in the CTO’s phrase, “logging in, not breaking in”: XSS leading to cross-tenant access with elevated privileges.

D2L’s position is that a similar attack on Brightspace would have failed and been detected quickly, based on two architectural claims: complete security isolation between tenants, with every support or integration action requiring independent per-tenant authentication, and real-time data egress monitoring. To its credit, the CTO paired the claim with appropriate nuance—“you cannot be smug about this”—and described ongoing hardening including reduced administrative access and AI-assisted code scanning. The analysis remains internal; customers are asking for the findings, and D2L has not decided on a sharing approach.

My read: D2L put more emphasis on the “we are more secure” direct argument during analyst day than in main sessions, and the descriptions were credible. I had an offline discussion, getting more in-depth descriptions, and there is real understanding behind the team’s leadership. But there are two caveats—in this age a cybersecurity hack could occur in any system, meaning that the incident response is at least as important as architecture; and poor messaging, tone, or timing can undercut the actual claims.

As I step back from the individual announcements, I saw much more coherence in strategy. Core Brightspace got unglamorous, daily-friction investment with dates attached. The announced Learner Mode is perhaps the most pedagogically explicit answer to the ChatGPT question (although still a bet). The content strategy is cleaner across native tools through Creator+ and H5P up to CreateSpace, with services added on top. And the ecosystem work should make Brightspace easier to connect to, not just harder to leave.

Which brings me back to where I started. With the most strategic product and services plan of the major LMS vendors, D2L does not need inflated win-rate claims or too-aggressive customer contacts to make that case. In fact, those instincts can invite skepticism about a story that increasingly stands up on its own.

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

Thanks for being a subscriber.