Instructure's analysis of 19.5 million higher ed users shows AI tools barely register in the learning stack and only 11 percent of instructors have comprehensive AI training. See the opportunity.

Only 11 Percent of Higher Ed Instructors Have Comprehensive AI Training, and a 19.5 Million User Dataset Shows Where the Opportunity Is

Instructure's first analysis of the higher education learning stack found that dedicated AI tools barely register in the data. Colleges have room to grow, and vendors have a clear picture of where the work is.

Key Takeaways

•  Instructure analyzed de-identified data from 19.5 million U.S. higher education users from September 2025 to April 2026 for its first look at higher ed LTI usage.

•  The most-used tools were McGraw Hill (11.4 million users), Watermark Course Evaluations & Surveys (10.6 million), Zoom (9.1 million), Kaltura (8.6 million), and Panopto (8.3 million).

•  Google Gemini, the most widely adopted dedicated AI tool, did not make the top 100, reaching 87,788 users across 211 institutions.

•  Only 11 percent of higher ed instructors have received comprehensive AI training, according to separate Instructure data cited in the article.

A New Look at What Colleges Actually Use

Conversations about artificial intelligence in higher education tend to be dominated by headlines and pilots. A new report offers something different: a look at what tools students and faculty actually touch. On September 30, Inside Higher Ed's Kathryn Palmer reported on Instructure's "EdTech Top 40: Inside the U.S. higher education digital learning stack," the company's first analysis of higher education usage of Learning Tools Interoperability integrations, or LTIs. These are external tools that connect to Canvas or another learning management system.

The analysis covers de-identified data from 19.5 million U.S. higher education users from September 2025 to April 2026. Instructure, which owns Canvas, had previously ranked LTI usage only in K-12. The report gives college leaders and vendors a rare, data-based view of the digital learning stack.

The Top of the Stack

The most-used tools are familiar names. McGraw Hill leads with 11,369,232 users, followed by Watermark Course Evaluations & Surveys with 10,632,703, Zoom with 9,051,162, Kaltura with 8,592,923, and Panopto with 8,333,378. About 90 percent of the top 40 were built specifically for higher education customers, with examples including McGraw Hill, Simple Syllabus, and MyOpenMath.

Roughly half of the top 40 are for content, collaboration, or video. Fifteen percent handle administrative functions and 12.5 percent aid productivity. In short, the stack that students and faculty use every day is built around content, communication, and administration, and those categories are where the largest audiences already are.

Where AI Shows Up, and Where It Does Not

Here is the surprise. Google Gemini, the most widely adopted dedicated AI tool in the data, did not make the top 100. It reached 87,788 users across 211 institutions. Ryan Lufkin, Instructure's vice president of global learning strategy, said he had expected ChatGPT, Claude, Gemini, and Copilot to top the list, and that institutions are not yet leveraging the LTI tools to embed them into the class.

The article is careful to note an important limitation. Lufkin pointed out that most faculty and students still interact with AI chatbots directly, which the LTI data does not capture. The data measures integrations into Canvas, not standalone AI use, so it should not be read as proof that few people use AI. It says that few institutions have embedded AI tools into their courses in a structured way.

The Training Gap

The most actionable number may be a separate one. Instructure data cited in the article found that only 11 percent of higher education instructors have received comprehensive AI training. Mary Styers, the company's director of evidence and learning strategy, said that integrating AI requires a higher bar than individual pilots or experimentation.

That is a clear opening. Training for faculty, support from centers for teaching and learning, and clear institutional guidelines are the foundation any AI integration will need. Colleges that invest here early will be better prepared when the tools mature, and the vendors, consultants, and training providers who help them have a concrete problem to solve.

A Note on the Source

Because this is a vendor's data, readers should keep context in mind. Instructure owns Canvas, so the report reflects Canvas integrations. The article does not give the number of institutions in the sample or a full methodology. Even so, a dataset covering 19.5 million users is a valuable window into one major platform, and its main message, that dedicated AI integrations are still early, matches what many campus leaders say.

How This Connects to Other Higher Ed Trends

The findings fit a broader pattern of cautious, deliberate adoption. Colleges are moving from pilots toward institution-wide strategies, as we have seen in system-level AI funding requests and campus-wide agreements, and they are expanding the programs that lead directly to jobs. Our look at Utah's technical college waitlist shows the same theme from another angle: where demand is clear, colleges and states are finding ways to fund it.

The learning stack is also where many of those decisions meet. A president or provost may set the strategy, but the chief information officer, instructional technology director, and learning management system administrator determine what runs on campus.

What the Top 40 Tells Vendors

The ranking has lessons for anyone building or selling to higher education. Tools that serve a daily need, such as video, course evaluations, and content, reach tens of millions of users because they are woven into how courses run. Tools built specifically for higher education dominate the list, which suggests that institutions value products that understand their workflows, accreditation requirements, and academic calendars.

For new entrants, including AI vendors, the path to scale is likely to run through integration. A tool that connects cleanly to the learning management system, respects privacy requirements, and fits into the way faculty already teach has a better chance than one that asks everyone to adopt a new workflow. The data suggests the market is open, and that the work of embedding AI is still ahead.

Building a Faculty AI Training Plan

Institutions that want to move beyond the 11 percent figure can start with a practical plan. First, define what comprehensive training means: not a single webinar, but a sequence that covers how the tools work, how to design assignments, how to discuss academic integrity, and how to protect student data. Second, offer it through the center for teaching and learning, with options for different disciplines.

Third, recognize the time faculty invest, through stipends, course releases, or credit toward promotion and tenure. Fourth, share examples from colleagues, since faculty often trust peers more than vendors. Fifth, gather feedback and publish what works. A visible, well-supported program turns scattered pilots into an institutional capacity.

Questions Presidents and Provosts Are Asking

Leaders are asking a familiar set of questions. Which tools are being used on our campus, and by how many people? Which of them support teaching and learning goals? How do we know whether the tools are improving outcomes? What does our contract with the learning management system vendor allow us to integrate? What do students and faculty need to feel confident?

Usage analytics like those in the Instructure report are exactly what answers the first question. Institutions that gather their own data, on tools, training, and outcomes, will make better decisions about renewals, pilots, and new purchases, and will be able to explain those decisions to boards and accreditors.

Signals to Watch Through the Year

Several developments are worth following: whether future editions of the Top 40 show AI tools climbing the ranking, how many institutions publish formal AI training programs, how vendors respond with integrations that fit the learning stack, and how accreditors and state systems treat AI in teaching. Each will show how quickly the gap between hype and embedded practice closes.

Who Buys Educational Technology at a University

The buying team is broader than many vendors expect. Chief information officers and provosts set priorities and budgets. Instructional technology directors and academic technology directors evaluate tools. Centers for teaching and learning shape faculty adoption and training. Online learning directors manage programs that depend on the stack, and learning management system administrators handle integrations.

College Leads' FAQ explains how its lists work if you are mapping a buying team. In practice, a successful campaign reaches both the strategic buyer and the person who will run the tool.

What Students Gain From Better Integration

The point of embedding tools in the learning management system is not novelty. It is convenience and consistency. A student who finds a tutoring assistant, a writing aid, or a study tool inside the course they are already taking does not need another login or another set of rules. An instructor who can see how tools are used in their own course can adjust assignments and guidance accordingly.

That is also why training comes first. A tool placed inside a course without guidance can produce confusion, uneven use, and integrity concerns. A tool introduced with clear expectations, examples, and support can become part of how a course works. Institutions that pair integration with training are likely to see the strongest results.

Privacy, Accessibility, and Procurement

Any tool that connects to the learning management system touches student data, so procurement teams will ask about privacy, security, and accessibility. Vendors that arrive with clear documentation, including data handling, accessibility conformance, and contract terms, move through reviews faster. Institutions that have standard review checklists for new integrations reduce risk and speed adoption.

The 19.5 million user dataset is a reminder of the scale involved. When a tool reaches millions of users, small details about privacy and accessibility matter a great deal. That is true for established products and even more true for new AI tools that are still earning trust.

Lessons From the Winners

The tools at the top of the ranking share traits that newer products can learn from. They solve a recurring problem, they work inside existing workflows, they are supported by institutional contracts, and they are backed by training and service teams. Reaching 11 million users is not an accident. It reflects years of building relationships with the people who decide what runs on campus.

What Vendors and Partners Should Do Next

For companies in this market, the data suggests several opportunities. Training and professional development for faculty is the most direct. Tools that integrate cleanly into the learning management system, with clear privacy and accessibility documentation, address the embedding gap. Analytics that show which tools are actually used help leaders make decisions with evidence.

College Leads maintains college email lists, a higher education contact list, and university mailing lists built around the buying committee, including a chief information officer email list, a provost email list, an instructional technology director email list, an academic technology director email list, a center for teaching and learning director email list, a learning management system administrator list, and an online learning director email list. These are the higher ed decision makers who will shape the learning stack this year. If you are ready to buy college email list data, or want a higher education marketing database filtered by role and institution type, build your list here and reach them while technology plans are taking shape.

Frequently Asked Questions

How many college instructors have AI training?

Instructure data cited by Inside Higher Ed found that only 11 percent of higher education instructors have received comprehensive AI training.

What are the most used learning tools in higher education?

In Instructure's analysis of 19.5 million U.S. higher education users, the most-used integrations were McGraw Hill (11,369,232 users), Watermark Course Evaluations & Surveys (10,632,703), Zoom (9,051,162), Kaltura (8,592,923), and Panopto (8,333,378).

Are colleges using AI tools inside the learning management system?

Not widely. Google Gemini, the most widely adopted dedicated AI tool in the data, reached 87,788 users across 211 institutions and did not make the top 100, though the data does not capture standalone chatbot use outside Canvas.