Moving from Canvas Core to Canvas Next

Hi, everyone. Good morning. Welcome. We'll give just a couple of minutes here before we get started so we can give everyone a proper time to join. Once again, we're just gonna give, a little bit of time here so more people can join us, but welcome again. And then while we wait, if you join us for the first webinar in this series, welcome back.

It's really great to see you. Alright. I'm gonna I'm seeing more people joining us. Thank you so much for being here with us today, especially if you're joining us over lunchtime. We're just gonna give a couple of seconds more here before we start.

And, again, today, we're continuing the conversation that we have done in the first session, but this time looking more at Canvas Next. If this is your first session with us, you can see the recording of the first webinar through the link that we will add in the chat. But just be sure that we also give you the context around core, around Canvas plus before we actually go into Canvas Next. And just a reminder as well, you can send questions throughout the questions and answers that is here in the in the chat. We have kept the final fifteen minutes specifically for that part of the conversation.

So I'm just gonna wait a little bit more here, and we're gonna kick off. Alright. Three minutes is my limit here. Thank you again for joining us today. And, again, this is the second conversation in our Canvas Tears webinar series.

In the first session, we have a look at Canvas plus, some of the ways that institutions can create, can have more visibility across the learning environment, can support people closer to where they are working and where they are learning, and also on sustaining more teaching practices across a much larger number of courses as well as users. So today, we are going to extend that conversation. Staff and learners are already using AI for different parts of teaching, different parts of learning and administration, while many everyday process is severely depend on moving between information, between, course screens, separate tools, and manual steps as well. That creates an opportunity for us and for all the institutions that work with us to think about where some of those activities can come together more naturally inside the learning environment with clear choices around how AI is introduced and used by the educators, by the learners, and by the administrators as well, many of you who are with us today. And that space is where we'll be exploring through in Canvas Next.

Just like last time, we're gonna spend a good part of the session actually showing you the experience from different perspectives instead of, going through forty five minutes showing slides and talking about, that those products inside of the slides. And for that, we'll look at the administrator, the educator, and the learner experience. And we have kept the final fifteen minutes open for your questions as Georgina also mentioned here in the chat right now. So before we get into that, let me just quickly introduce myself and the team. So for those who I haven't met yet, I'm Joao, and I look after the market strategy for EMEA.

And it's a pleasure to see all of you today. I'm seeing some familiar names in the list, so thank you so much for being here. And I'm here with my good friends, Andre and Fran. Andre, would you like to say hello to everyone and explain a bit of what you do today? Hello, everyone. Andre Silva here.

So I'm a senior solutions engineer here at Instructure. My academic background is in education and professional background in digital transformation, helping companies, institutions moving from more traditional approaches to education to more digitally enabled, so that they can leverage technologies. So today, we're gonna be covering some technicals, but I cover a whole range of technical aspects and pedagogical considerations as well. Awesome. Thanks so much, Andre.

Andre is our demo ninja, so I'm happy that he is here with us today. And we also have the one and only Fran. Welcome back, and thank you for being here with us. Would you like to introduce yourself to the group? Of course. Thank you.

Hey, everyone. Good to see you. I am Fran. I'm from the success team here at Instructure, and my job is to essentially talk to a lot of you on a regular basis and really understand the pain points, the usage, and, everything that you could benefit from, with you know, within the Canvas ecosystem. So I will I will bring a bit of a, of a kind of a success flavor to the conversation.

Alright. So you're you're gonna hear from the three of us as we move between the educational context, the customer experience, and the product itself as you can see. But before I say anything in addition about Canvas Next, we we wanna hear a little bit from you. So let me ask you a question in the next slide. And we are gonna start with a quick poll here.

Think about, you know, what is taking most time, what is creating the most discussions, what is requiring the most coordinating your institution is, for example, the administrative workload around managing the learning environment. Is it helping staff to get the data that they need and work with it more easily? It could be also supporting learners as AI becomes a more common part of how they study and for for what do they study. Or is it making sure that new technologies introduce it consistently with clear expectations around it, around the use itself of them. So choose one that feels more close to what your institution is working through at the moment, and we are gonna see the results in in a few seconds. Interesting.

Alright. I think we have a good number here. While we are closing the the the results here, there are two of those areas that I want to keep in mind as we go through this session. One is, the amount of time that people spend moving from information into action. You might already know that something needs attention, but investigating it, understanding that context, and then following through can still involve several people in several steps.

Right? And the other one is AI use itself. This is one of the the hottest topics in everywhere right now. So while institutions are increasingly deciding what kind of experiences they want to provide within their own learning environment, there's still a lot of room to create governance the experiences between the educators, the learners, and the administrators. So those two threads will come together several times today with also, the things that you have just entered inside of the pool. And the place that we need to start while we talk about Canvas next really is the foundation underneath those solutions.

So let's take a look on what that means in the next slide. Canvas core is really the foundation that keeps moving. So for those of you who are already using Canvas LMS, the core is that foundation that is supporting a significant part of your everyday teaching and learning experience. But for those of you who might be evaluating Canvas, this is really the starting point in the tier structure, Canvas Core, Canvas Plus, Canvas Next. So Canvas Core brings together that central experience around the courses, assessment, feedback, accessibility, communication.

And that foundation, as I mentioned, continues to involve, including improvements to teaching workflows as well as Ignite AI capabilities that are available within Canvas Core. So Core is really the strong option for those institutions who prioritize and who the priority is really having a reliable environment for everyday teaching and learning. But the thing is that as different needs develop around that foundation, institutions can then decide whether additional capabilities in Plus or in Next are relevant to what they are trying to achieve. And that distinction is worth making briefly before we go any further. So let me show you in the next slide, how can we connect that foundation in the ways of working.

Again, if you join us on Canvas plus webinar, this will look familiar to you. Core, as I mentioned before, is the everyday teaching and learning foundation. Canvas plus adds capabilities around visibility, support, around engagement, the interactive learning, and teaching consistency, a lot of the things that, Andre has shared with us and demoed in the last week. So for example, an institution can get clear signals across learning activity, support people within the context where they need help, and give educators additional support around teaching and feedback. But Canvas Next includes all of that that is in Canvas Core and Canvas Plus experiences while extending what people can do around data, around the Canvas workflows, and the learner support itself.

So I would say that the distinction between what you can get with Canvas plus and with Canvas Next is actually really important for you to to understand. The visibility and insights we discussed in the last session really continue to matter, but with Canvas Next, people can take some of those interactions further by asking more specific questions, working through tasks in different ways, bringing an AI supported study closer to the course itself. And that becomes very relevant as the way people are using AI starts developing faster than some of the processes around it. Right? In fact, if we look at, some numbers that came from the Digital Education Council recently on the AI Higher Education Global Service that, survey that they've done in 2026, so this year, use AI in their teaching in more than 35 countries that were, that the service survey went through. But then only 31% of the faculty members feels that institutions are meaningfully involving them on the AI policy.

So what I we wanna show you next is really not on not only on the next slide, but really on on this webinar is giving you a closer look on what does that mean and how institutions are working to reduce that gap. So, really, what happens when AI adoption moves faster than the systems around it? You can see this in really fairly ordinary situations. Right? Like, an educator might use one AI tool while preparing teaching materials, but a learner might use another one while revising, while studying, or even answering assessments. Let's not even start on which models of the LLM are people using or even how they are prompting. Right? So there's a lot there.

And at the same time, the institutions may be still deciding or in the process of deciding which tools they actually want to support when it comes to AI, what informations people should, share with them, what is the guidance that should be in place. So those same decisions become more significant as use becomes more common across different parts of the organization. And the opportunity comes from bringing more of that activity into the learning environment they already use and giving institutions a clear way to decide how, and which of those experiences should work for the learners, the educators, and the administrators as well. Fran, learner behavior, I think, gives us a useful indication of how established some of this already is. Right? It certainly does.

Thanks, Joao. If we look on the next slide, we really look beyond a, single institution. We can just see how established these behaviors are and have already become. So Eurostat found that almost two thirds of sixteen to twenty four year olds across the European Union, used generative AI in 2025, and around 39% of them were already using it for formal education, and we can only assume that that's gone up in So if we then look specifically at higher education in The UK, the HEPI data goes even further. So ninety five percent of the undergrad undergraduates that surveyed said that the use of AI is there in at least one way.

So at the same time, we have this, and we know also that only 38% said that their institution provides them with AI tools. So these are different studies and different populations. So I wouldn't necessarily compare the percentages directly, but what they show together is how established AI use already is amongst learners. And for the institutions, that creates some, very practical decisions and attrition in some ways. So what kind of AI supported, experience do you actually want to provide, and what should that experience be grounded in? What sorts of guidance should learners have around it? And where do you want the institutions to have a clearer oversight of how those experiences are designed and introduced? So you can give learners an institution supported option inside the learning environment grounded in the content that they're actually studying with clearer parameters around how that experience works.

So that brings us directly into what Canvas Next is designed to add, really. So, obviously, from a customer experience, a customer success perspective, this is a massive operational vulnerability as you can imagine. So if institutions do not provide a safe governed AI ecosystem, learners and staff will inevitably resort to the the Wild West of public unvetted AI tools, and they already have. So this this means that your proprietary course data and student data are leaving your secure environment. And what we are doing within Canvas Next is to give you the power back to bring that activity back inside the learning environment where you already manage and govern and protect your data while still driving that innovation that comes with the use of AI.

And that is quite a a useful way into what Canvas Next is designed for, isn't it, Ujjwal? Yeah. Absolutely. And I love the analogy there on the western. There are three main ways, I would say, that you will see in the next slide in terms of for the purposes of Canvas Next. The first one is really understanding.

Right? Teams can already see useful signals across the learning environment, but Canvas Next really gives them a more direct way to ask follow-up questions of the data in natural language and explore what sits behind what they're seeing. Then there's action, both for educators and administrators. They often know what needs to happen in Canvas. Right? But carrying it out can still means working through several separate steps. Show Canvas Next really gives them a way to describe the outcomes they want in natural language and carry out their work across Canvas while they will always stay and continue to review and guide what happens as well as approve.

And then there's a learning piece here. Learners can work directly within the course content they already are studying and use it to practice, check their understanding, and reinforce what they've learned without having to recreate that context somewhere else. So across three areas, Canvas Next really changes how people can work with information, how they can move from intent into action, and how learners can work with the material already in their course. Those experiences, they are supported through EscrowData, Ignite AI agent, Ignite AI study tools. And Android is also going to show you, what each of them looks like in practice plus an additional capability that is under development right now, for our learners.

And because this capability sit within Canvas, institutions also, as under, Andre and Fran, were talking before, have a clear place to decide how they want to introduce them, who should have access, what expectations should should sit around their use, which I believe is the most important piece here in the support of a common use of AI in your institution. So let's say, for example, in the next slide, a reporting example. A CIO may already have a clear view of learning engagement across institution, but, really, the next questions are usually more usually more specific. Right? Which courses are driving that pattern? Does it look different across a particular group of students? That's where the workflow can slow down because answering those questions may still require another report support from the data team. So, Andre, so that I can stop talking a little bit here, can you show us how Ask your Data changes that experience, please? Absolutely.

And we will start with Analytics Hub. And before we jump into the demo, just to be, just let's just reflect upon where we are, in terms of status quo for data and analytics in learning management systems. So we are coming from a world where the industry standard is for platforms, learning platforms, to provide their clients with the ability to see a number of fixed or somewhat customizable, visualizations and data reports, but really within set parameters. So really, what we have been so far or what we've witnessed so far is that you have a bit of a normal distribution where, people, the most use cases can be supported. We can create visualizations.

We can create dashboards, the the industry as a whole. But then it's very difficult to service particular requirements that sit on the tails that are not, as frequent. Okay? And what we have done what we're gonna be, showing you today is how SQL data can take you to the next level. It's just a point. If you wanted to create more customizable dashboards, you could use could extract data tables.

You could, maintain, data data lakes or data warehouses. We'd have to have licenses with third party tools, for example, Tableau or Power BI, where we could create more tailored, dashboards. And yet that was possible, but some institutions that were not as technically able or could not invest as much would not be able to access, to that information in those particular dashboards. So what AscioData is gonna be offering you today is the ability to create customized visualizations at scale for your particular institution. So what I'll do is I'll go into data assisted insights.

So by the way, this lives inside your, analytics hub. And in this case, what I'm going to show you is it's very simple for you to create dashboards, in this case KPIs, multiple dashboards. Dashboards can be, for example, per division, can be for staff, can be for learners. And here, what you can do is each one of the visualizations created, are relevant to you, to a particular use case you have in your institution. So no longer that approach, normal distribution where we serve most use cases, but then you feel like actually the use cases that are relevant to you are not covered in the platform.

So here what you can see is, for example, I've set a KPI, where a value is compared against a particular target. Here, you can see I have visualizations that are showing me, for example, how my institution as a whole is using, or courses in my institution are using particular, resources. Here, you can see the number of pending assignments and quizzes. So all of these I've created because they saved my particular use case. So the question is not that just you can create these visualizations or service these data, in some cases some reports, directly in the dashboard.

But we have simplified the way that you can create these dashboards for your particular use case. And what I'll do is for this particular just using using this as as an example, I'm going to jump into the chat, and you'll see that the way to create visualizations is via natural language processing. You can just ask a question like I've I've done here, and the ask your data is gonna be going through the system. And in this case, the one source of truth, the the data access platform, which has data of all the things that happen in Canvas. It understands the data tables and it understands your query and is going to be trying to marry the two to provide you the answer you want.

And in this case, you can see that I've raised the first query and surfaced some data on the right. You can tell you can tell or you can see here what has been done. So there's a summary of what has been done in terms of, data analysis and what has been found and the methodology that has been used. In some cases, there may be assumptions. We humans use some words, interchangeably, and what AI is gonna be trying to do is understand exactly what those terms are and link those two terms that are used in our data tables.

Okay? In this case, I was not happy with the first query, so I expanded and asked a new question for for for added addition of a new column, which AI did. And what happened was, as a new table was surfaced with the data that I have or data sets or in this case the the column that I had requested. Once you have a data set, you can start creating visualization. So you can see edit chart and we we have quite it's super simple to use, but at the same time it's fully featured. So in this case, I can set up what are gonna be the categories that are gonna be used.

In this case, it knows that this I can look at this from multiple angles, angles, course name or learner name, and year count. So I want to see from all my courses what are the courses that have the highest number of unresponded, or, unsubmitted, assignments. So you can see here, I want to count, so that brings everything together. And you can, of course, change, the visualization to use a visualization that is applicable to you. In this case, I finished.

I can now, look at the data SQL the the SQL prompt. So in the background, what AI has done for you is created a SQL prompt. You can edit this. And most excitingly, you can, look at the diagram, the data diagram that shows how different tables are interacting with each other. Now, let me just move this to the side.

Close here. Once you are happy, in this case, I'm gonna be sharing a chart. So I'm going to use the chart instead. Oops. So just save the chart.

I'm going to use the chart form rather than the report form, and I can now pin it to a particular dashboard, as you can see here. So as you can see, this is transformational. We are enabling you to first create dashboards that are relevant to you, to your particular use cases, in only your particular use cases, right? And we are giving you the tools so that you can start cut through the complexity and collapse layers of complexity, additional licenses, external licenses into something that is can be created, maintained inside the system cohesively. And, of course, what you see in the end is a is a is the pin board. In this case, I could add, this particular, chart to a particular, dashboard or to multiple dashboards, and then you can share that widely with your teams.

Okay? So to close, Fran, over to you. Thank you, Andre. That was that was brilliant. Yeah. What I would take from this brief demo is how much of the existing process can sit with, the person who is already trying to understand the situation.

A head of school or a program lead may understand the educational question extremely well and still depend on somebody else to translate that question into a technical query. So here, they can explore the question themselves, see how the answer was produced, and turn the results into something that they can discuss with with colleagues. So there will be still a deeper analytical work that belongs with specialist teams. However, this gives much more to people a direct and democratic way to investigate the questions that naturally come in in in areas that they're responsible for. And with that, when we move from understanding something to actually changing something inside Canvas, that takes us into a different capability.

And, Joao, you're going to tell us more about it. Yeah. After we think about everything that is in regards to the administrative side of seeing the data, like an educator may already know the outcome they want, but getting there can still mean several separate actions across the course. This could be, for example, and it's on the next couple of slides. There's a question there that I think it's really interesting when we think about the the educator point of view here is how much teaching time disappears into the course management for our educators.

Right? There's a lot of great things that are already available to support the educators inside of Canvas Core as well as in Canvas Plus. But when it comes to, Canvas Next, there's other ways that can even better support review engagement, adding support materials, and then communicating with the learners who need that attention. I think that overall, even more is being able to reduce the amount of time that is spent into those day to day activities that are time consuming in the administrative point of view for the educators so that they can focus on the next step of the learning for their students. So, Andre, can you show us how agent helps us to bring those steps altogether? Absolutely. And I will jump on to my teacher.

So in this case, I'm in my teaching account. And I I would say, in its simplest form, the Ignite AI agent is there to help you. So if you are starting to use the system and you know exactly where to go, where to press, or how to do things, you can just use it as a as a simple tool to ask questions. However, if you are thinking about driving the maximum value and what it can actually do for you, well, exactly as John was mentioning, we can use it to create multiple stepped processes. And but first, what does he actually do? So what in the background is happening is the agent enables you to ask questions and perform actions in the course.

So in the simplest form, he understands your queries via natural language processing, and he's gonna be interacting with APIs in the background to trigger particular actions for you. K? And we're talking about things that we can do for you, course management, student management, communication communication with learners, and and assessment. So course management, creating organized modules, for example, can do that for you. Student management can see student progress or engagement, for example. Communication can create inteler communications for you, announcements, for learners, for example.

And then you can create assignments and track submissions and grades. So with that said, for example, to show you what what is possible here, I have a simple simple example where I have actually two requests in one. So I want to in in understand learners' engagement. So in this case, I'm I'm specifically saying learning learners' engagement is gonna be tracked by less participation in discussions, greater than five days, fewer posts than average. So I'm defining specifically what I want the agent to consider as as engagement indicators.

And then based on those, I want to look at learners that are at risk based on those indicators. And then, and based on the the lowest scores, try to create content for these learners. So propose materials that it can create for us for those particular learners that are in it. And then finally, write a notification. So we have three steps here, and you'll see that if I ask a particular this particular question, the Ignite AI agent is gonna be understanding what I'm trying to do, and he's gonna start immediately looking at the learners, trying to get identify the learners at risk, and what he's gonna be doing after that is creating a plan to create content.

And you'll see that I am asking to wait for my approval for content, so I don't want to just just go and and create content. I want to have this human in the loop where before any actions and changes are made in the course, I want you to interact with us. And, also, I'm not gonna be completing this complete flow. But at the end, what would typically happen is once a proposal for content has been surfaced. So it's gonna be surfacing learners at risk and lower engagement.

It's gonna be suggesting topics or content that can be created to support those learners in difficulty. And then what would happen is, yes, we'd create content, and then we'll create craft a notification for learners and ask me, are you happy with that announcement? Can I send these to the learners? And then you will always have this approval before any communication is sent to learners. So in this case, let's have a look very quickly before I hand over to to Fran just to see what the agent has done for us. So as you identify the students at risk based on a number of indicators. Right? And based on those learners at risk summary, that give give me a summary, which learners are the ones of highest concern, and then areas of improvement that has those particular learners identified areas with knowledge gaps, and then he's proposing additional resources for those particular learners at R, that have, those particular knowledge gaps and how we can bridge those knowledge gaps.

And I would imagine at the end, are you happy for me to continue? Now, at this point, you could add the file, for example, if you want to to use that as a baseline for materials. But this gives you a sense of how I've created or asked for a quite advanced, quite complex, three step process. And you have this Ignite agent helping you, guiding you through what's needed, and asking you inputs only when necessary. Okay? So, Fran, over to you. Thank you, Andre.

This is by far my favorite, feature that we have currently. Really, really good. What I love particularly about it is that from an adoption standpoint is that we are enforcing you and your educators to learn a completely new disconnected software. Right now, we have instructors spending hours, cross referencing engagement scores and identifying at risk students to get that kind of differentiation in place that we so need now and and really manually searching for remediation materials and copying and pasting individual emails. And with the agent, you are effectively giving every educator a highly efficient teaching assistant that operates across 500 plus Canvas APIs.

And I think that's that's that's quite it's quite cool. The educator simply states the intent, like you've said, like, yeah, you've seen just now with with Andre. For example, like, identifying students who failed the last module and drafting a supportive message, and the agent orchestrates a lot of that heavy lifting. So this is particularly important so that the educator can stay fully in the loop to review and approve the action. So it's never a replace.

They're never replaced, but it's about preserving academic judgment, while completely eliminating that administrative fatigue, that is is a big struggle for a lot of, educators at this moment in time. But that's obviously, all from the educator's perspective. What about the students, Gerard? Yeah. So when we think about the students and looking at that same digital educational council survey for from more and more the usage of AI on the learners. In fact, 88% of the learners who went through that survey mentioned that they are using AI on their day to day learning.

So I would say that while in 2023, '24, a little bit on 2025, the question was how do we stop students from cheating with AI? We actually are moving to what does it mean to teach and to learn when AI is actually a permanent part of the room. Right? A learner may want to check their understanding or practice before an assessment. And today, that often means taking course content into a separate AI tool. But with Canvas Next, our study tools keep that experience and the evidence connected to the materials that they are already working with. And Nandra will show you how that works, but then also give a short preview of knowledge chats, which is a unreleased capability that we are exploring, that we are developing, that we have sat down with students here from the EMEA region as well, but is a more conversational learning support.

So, Andrey, I'll I'll leave it to you. Thank you, Juan. So what I'll do is I'll jump on to the learning site, our learning site. I'm going to go to modules, so that I can see my sequence of learning steps that I have to complete. And in this case, I'm going to go to a native page.

And what you'll see here is that on the top right, I have this Ignite study Ignite AI study tools. Well, I'm gonna click on it, and you can see that I have three options. I can summarize content. So it's summarizing content of this page. This works as well with PDF files and docs.

You can also access quiz me feature. The quiz me will, based on the content, generate a number of questions for you. And you can interact with these quizzes directly here, and will generate 10 questions. So once you have run out of questions, you can regenerate and create more. By the way, this is gonna be formative assessment rather than summative assessment, but it's this idea that we're enabling learners to approach content from multiple angles.

And then finally, you have the flashcards. Flashcards, very much like a quiz, will ask you a question, and then you can turn it and see the answer. Alright? So on its own, it offers you features to exactly try to understand the content a bit better, summarize it, ask you questions on the fly, at at level of interactivity. I would say this starts to be very interesting when you when you start linking these, for example, with the notebook. So now you have identified a few areas that you have difficulties or that you found important because of the summaries, because of the quizzes, and can mark as unclear, for example, and add a note.

By the way, this notebook has been added as part of our system, but I just wanted to show you how can be an interplay between these AI components and features that we are building around it as well. All right? Now, to talk about, this is already available, study tools. The next feature I'm going to show you is going to give you a bit of an insight of what's about to come. And I wanted to show you because this is where we take education to the next level. And you can see here knowledge chats.

So knowledge chats enable me to engage with a bit of a coach, a digital coach. And you can see here, I can click I have a number of conversations for this particular course. I can click on this particular conversation, and you can see that the coach or AI coach has asked me a question and based on that question there has been a back and forth between me and it. And you can see from the institution side, which I have here, give me one second, We have created this. We are on the side of the creation of this, particular knowledge chat.

I just wanted to show you this because it show you the guardrails and, the the the the thinking behind that particular component. For this particular knowledge chat, we have defined what the learning objectives are, what the pedagogical activity or guidance is going to be for this particular AI component. So what is the personality it's going to take? What particular characteristics? If you want to if you have any nuances for this interaction, here's where you define that. And down down here, have text source. So basically, we are giving a body of knowledge that this component is going to be basing itself on.

And you can as well just gonna add it so that you can see. You can also add files in addition to the knowledge base or knowledge body. I can see here, you can upload a file. And basically, what happens is you are creating the guardrails through which this interaction is gonna be taking place. Okay? Now going back to the learner, as I interact with the chat coach, you can see that the learning objectives that you have set as an institution start to be tagged or ticked automatically.

And you can see that this is happening basically via conversation back and forth between the the coach and you. And once you have completed all of all of the learning outcomes, you would have completed this chat. Okay? So at the simplest form, this can be a way to check that the learner is hitting particular knowledge points or is meeting a particular threshold of knowledge. But if you start thinking a bit a bit more widely, this can also be used as a coaching process or a coaching component as learners are, for example, applying skills on a particular setting. In this case, this learner is discussing with the the coach some things that have been happening at work.

And you can see there's been questioning, are you applying the right skills? Are you applying the right principles? And this back and forth can be very healthy. Okay? Finally, back to the institution front. If you are concerned about and some of you by the way, all of these aspects, these AI components we have been covering today, they are optional. So you can implement the ones that you think are appropriate to you. But in this case, if you are concerned about these interactions with AI and you want to have full insight of what what what is happening between the learners and and and this knowledge chat component, you can come here and see all your learners in your class, and you can see the discussions from the teacher's side.

You can see the interactions that have happened and that have taken place. Okay? So if there are any concerns, can immediately turn turn it off or make adjustments as necessary. Okay? So we are what we're talking about here is giving you a component with very tight guardrails and controls that enable you to scale up this level of one to one interactions that in the past would always have to be done with a human. Now you can really have tight controls on this coaching aspect and deliver this at scale while still keeping control. Okay? Right.

So this covers the the two features I wanted to cover with you today, and and it over to you, Frank. Thank you, Andre. Yes. Again. Yeah.

I think the combination of those two components, the study tools and the knowledge charts is really super useful. They directly solve the academic integrity and the privacy conversation that is so, so relevant at the moment. And that one that I have with with our partners in Europe all the time. So, currently, if a student wants to practice for an exam, they are copying your institution's copyrighted lecture notes and pasting them into public AI chatbots to generate flashcards. With study tools, they can generate practice quizzes and flashcards directly from the Canvas content that they are already viewing.

And looking ahead to the knowledge charts, this really allows you to provide personalized one to one coaching at scale. Because the institution enhances the knowledge base, the AI isn't hallucinating answers from the open web, but it is actively guiding the student towards your specific learning outcomes. And that ensures that the AI serves the pedagogy and not the other way around. So for skills based learning, that could also create space for a learner to reflect on how they would apply what they've just learned in practice. And I just wanted to reiterate that what you just saw with the knowledge charts is still unreleased.

We just wanted to share it with you because it kind of helps explain the direction that where we were the direction that we were exploring around learner facing AI, which many of you has asked us to. So if we put that alongside ask your data and the agent, you can really start to see why the Canvas Next story extends across several different parts of the of the learning environment. Xiaowei, it's back to you. Yeah. A 100% with you on that, Fran.

And, because we are almost in the limit of the time here, I'm gonna be really quickly on this one. But I think as Fran was mentioning, this is where the three experiences become really just one story. Right? Canvas already holds a huge amount of the context that makes these experiences useful. It understands the courses, the people, the roles, the activity, understands the content, understands the outcomes, the permissions around them. So that context changes what AI can actually do inside the learning environment, and there's an interesting question on the q and a that goes into those lines here.

A question about data can start from the right institutional scope. An educator can act within the course they already manage. A learner experience can stay grounded in the content and the learning outcomes the institution has defined. This makes governance a much more concrete conversation for our institution. It shows up in who can access what, what information the AI is working from, what experiences is allowed to do, and where the person still owns the decision.

That's the connection that we wanna bring on everything that we have shown today, and that's the connection that Canvas Next brings with Ignite AI into the educational context that Canvas already understands from your institution. So, Fran, why move to Canvas Next? Of course. Why move? So I start with what is already happening in the institution, really. So you may have teams that are making a very good use of Canvas data and spending a lot of time dealing with follow-up question through a separate reporting process. You may have educators who know exactly what they want to change in a course and still spend a significant amount of time carrying out those repeated admin steps.

Or you may have learners already using AI heavily like we saw earlier for revision as your institution is still deciding what kind of study support it wants to provide within Canvas and through AI. Or you might already be introducing AI more broadly and want more of a of that activity to happen within the systems that you can you can manage now and you can trust already. So any of those can really be valid reasons to explore next. So the usual thing is being specific about what where you want things to to improve. And you can identify your recurring workflow, the people involved, and what a better experience would look like, then you can have something concrete to evaluate.

And, of course, these are the types of conversation that I that I have and that the team has on a regular basis. So that gives, I guess, that gives you a much stronger starting point than simply asking whether institutions want additional AI capabilities. And, if we go on to the next slide, similarly to what we suggested a couple of weeks ago for Canvas plus, for the Canvas plus webinar, moving to Canvas next does does not require, an LMS migration. So, you keep the Canvas environment you already use, and the next add ons, the additional capabilities, we've been looking at across data, AI supported workflows, and learner studies are just, enabled there. So the next conversation can start from one of the areas that you may have recognized as relevant for you today.

And and one of the questions that you might have of which workflow do you want to improve? Where where are educators, administrators, or support team spending the time that they could be used differently? And what would you like people to be able to do with the data that that they already have? Or from a learning's perspective that, you know, learner learners that are already using AI, what kind of experience would you like to provide to them within that Canvas environment? And from there, obviously, myself, if I'm your customer success manager or one of my colleagues and the bigger account team that you have can really help you identify which next workflows is worth exploring. And and, of course, we can involve Andre as we have done or someone in in Andre's team where product or technical validation is needed, and we can have a kind of a one to one conversation that is more linked to your real needs. So for for you who are still evaluating Canvas, the same questions can help you think about what what do you want from your learning environment? What do you want your learning environment to support as these ways of working become more and more common, really? So that is that is why why why next and what to do next, really. Thank you, Fran. And I guess you have an invitation as well to join the next slide.

I do indeed. So as always, we have our CanvasCon, which is our main event of the year here in here in EMEA and this time is in London. So please join us. You can see the QR code here. There is a page dedicated to our community in our community to registering for it.

Many of you have already registered, so really, really looking forward to seeing you there. Second and third of November, we're coming together, and we we expect more conversations, more workshops, and more updates. Who knows? But, yeah, do join us. Awesome. Thank you.

And this gives us ten minutes to questions and answers. So not the fifteen that we promised before. Apologies for that, but at least we had a good discussion here, the three of us. And we do have some interesting, questions here. Some more technical, some other even more philosophical as well, I would say, but let's start with the first one here.

Andre, I think this one is for you. Can Ask Your Data pull data from third party tools integrated on Canvas via API or LTI? So we would not be able to see what happens inside those tools because that data is not made available to our system. So Ask Your Data will be focused on the data tables and data available inside the system. Perfect. Thank you.

Then what if you don't wanna specify and I assume this is for agent. What if you don't wanna specify to wait for your approval when it's creating something for you? And I think the the approval perspective here is a really interesting one, not only from the technical point of view, right, but, you know, the reasons of why we are stimulating to to always ask for that approval. So I don't know if you wanna share something about it as well. Yeah. So I would say this is gonna be a use case by use case.

I would say there are specific workflows that will always ask you for a confirmation, for example, contacting emails. I have not tried, but I would assume if I would have asked to for for it to create content immediately for those learners, it would go ahead and do it. So then it's gonna be varying by workflow and use case specifically. However, I would say we are still early in the stages. We we I mean, we're talking about zeitgeist of implementation and use of these agents.

So we are adopting a conservative approach rather than allowing the agent just to go in and do things completely independently. We want to land these with a number of institutions, and we are still working on these and expanding features as we go, as we as we as we move along and learn as you learn. So I would say right now, things that are gonna be having an impact to learners will most likely ask you for a confirmation when in the future you and us can work together to define what the future is gonna be. Yeah. And I think this is, this is a good link to the next question here about human judgment, right, and the approval process, but also the the the fact that we are working to eliminate the administrative fatigue.

But what about introducing the verification fatigue? And this, to me, goes even beyond the conversations that we have here in in the development of our prost of our software and more into AI itself. Right? I think this is a really fair concern. We don't position AI as removing the need of the human judgment, specifically in higher stakes workflow. So the goal that we have right now is really to reduce the repetitive work while keeping people in control. So review is really focused where it matters, rather than adding another layer of checking everything and everywhere.

And then another question as well, really interesting here, is that schools have and this, I think, Andre, is more technical, so maybe more for you. But schools have different tool sets. Right? Not all Canvas features are used at all universities. So does agent adapt to this in some way? For example, the particular institution does not use, does not build in collaborations or using big blue buttons. So I think that overall and you can correct me here, but agent operates within the capabilities and permissions available in the Canvas environment.

Right? That that is correct. That is correct. And we'll know, the configuration that you currently have in your system. So we'll see, what are the features that that you are using in a particular course. So you can actually do these at the course level.

Yeah. But but, yes, he's gonna be aware. He's gonna be asking you questions about whether to interact with that particular component. Of course, he cannot perform LTI actions inside LTI, but we have already covered that. Yeah.

Although there's a there's an interesting also topic coming more and more here at Instructure, which I will leave to what London is calling you to go, which is for CanvasCon in Europe that is gonna happen in London. That relates to MCPs and structure strategy into that side of of things when it comes to AI and what we are building in that area. So keep an eye on CanvasCon because this is where we're gonna spend a lot of our time as well talking about this. Now the last question that we have here, and we still have five minutes, so please feel free to share more questions if you have, is regarding the study two. Can the content be in a file uploaded to the course? For example, a PDF, Android.

It can. So PDF and Word documents are not officially supported, but I've tested myself and and it works with those documents. I am not sure if we, over time, are gonna be, making it available for more documents, but it's working for those more the documents at the moment. So yes. Amazing.

Well, I wish that all our attendees' microphones were open as well so we could continue the the discussions here. There are some interesting questions, and I appreciate everyone for the courage and and the curiosity to ask those for that for us. And I think this is the end of our webinar series, but, really, thank you, Andre. Thank you, Fran, and thank you for all of all you who are here with us today. And the recording will be available soon as well as if you wanna see how the Canvas core to Canvas plus webinar was, just make sure to see the recording that is in our website.

Thank you so much. Thank you, Fran. Thank you, Andre. Thank you, Joao. Bye, everyone.