The Lambda Learning Blog

AI in Totara: what is included, what it costs, and where it stops

Written by Lambda Learning | Sep 22, 2026

If you are evaluating Totara LMS, the AI question usually comes up in the second meeting. Someone on the committee has seen a competitor demo an AI coach or a skills graph, and wants to know what Totara has in its place.

The short answer: Totara has a small, practical set of AI features aimed at the people who build and run learning, plus a machine learning recommendations engine that has been in the platform for some time. It does not have a learner-facing AI tutor, and it does not infer skills automatically. What it offers in exchange is control over whether AI runs at all, which account it runs on, and which content it can see.

This guide covers every AI capability Totara documents, how AI is switched on, what it costs, and the questions a regulated organisation should ask before enabling it.

Two kinds of AI in Totara

It helps to separate them, because they work very differently.

Think of it as the difference between a colleague who reads the room and a contractor you call in. The recommendations engine reads the room: it watches activity inside your own Totara site and suggests what to look at next, without sending anything out. The generative features are the contractor: when someone asks for a summary or an image, Totara sends a request to OpenAI and brings the answer back.

That distinction matters most when your privacy office asks where the data goes, so it is worth holding on to through the rest of this page.

The six AI capabilities Totara documents

These are Totara's own feature names. Each is optional and configurable, and all of them are managed from the Manage AI page in site administration.

AI form assistant

Works inside Totara's Weka content editor. A course author or a learner creating a resource can summarise existing text, generate a new block of text from a prompt, or generate an image without leaving the editor. Typical use: a first draft of a course description or introduction, which a person then reviews and changes.

AI image generation, standalone

The same image generation as a dedicated tool, available anywhere Totara supports an image upload, including course settings. Useful for giving a large catalogue a consistent visual look without a design request for each course.

AI knowledge check-in

A facilitator uploads a file and Totara generates an informal quiz from it, with auto-generated questions. It sits alongside resources and surveys in collaborative workspaces. It is designed for low-stakes reflection and spotting knowledge gaps, not for graded or compliance assessment. For certification and recertification, Totara's formal quiz and completion tools remain the right tools.

AI course tag recommendations

Totara reads the information held on a course and recommends tags to add to it, drawn from your existing tag list. It does not invent new tags, and where tags exist in several languages it suggests them in the user's preferred language. On a large catalogue, consistent tagging is what makes search and recommendations work, so this is a quieter feature with an outsized effect.

AI goal assistant

Part of Totara Perform. It helps learners create SMART goals by walking them through a series of guided questions and drafting a goal that is specific, measurable, achievable, relevant and time-bound, which they can edit or accept.

AI-powered recommendations

The recommendation engine behind the Recommended for you block in the learning catalogue. It supports courses, micro-learning, workspaces, and programs and certifications where self-enrolment is enabled. It draws on interaction data, user profile attributes, content metadata and what is trending across the site.

According to Totara, this engine processes data within the Totara environment and does not send it to an outside provider. Site administrators choose which model it uses, including whether user profile data is considered at all.

How AI is switched on

AI in Totara is opt-in, and more so than the phrase usually implies.

The generative features are not part of a standard Totara installation. They are delivered as separate plugins that your Totara partner installs, and they need an OpenAI API account before they will do anything. Once installed, an administrator enables individual features and sets role-based permissions, so you can, for example, give course authors the form assistant without giving it to every learner.

Totara is candid that these plugins are early-stage. Its developer documentation describes them as evolving features, delivered separately from the core product so they can keep pace with AI, and not built to the same development and support standards as the core platform. For most organisations that is a reasonable trade for content drafting and tagging. It is also a good reason to keep AI away from anything that decides a compliance outcome.

What AI costs to run in Totara

There is no AI credit meter and no per-action charge from Totara. The generative features run on your own OpenAI account, so you pay OpenAI directly at its published rates for the usage you generate.

Two qualifications, so the picture is complete:

  • The plugins are built for OpenAI. Totara states that integrations with other model providers, such as Anthropic, Gemini or Llama, can be built, but that is partner customisation work rather than a setting.
  • Anything beyond the packaged features, such as an AI learning assistant trained on your own course content, is a custom build with a partner, and priced as a project.

Compared with platforms that meter AI per generated lesson or per minute of video, this model tends to be cheaper for light use and more predictable to budget. It is less convenient if you want a large AI feature set working on day one.

Questions to settle before you enable AI

For a health system, a regulator, an association or a public body, these are worth answering in writing first.

  1. Which content leaves the platform? Course tag recommendations send the information held on the course. The knowledge check-in generates questions from a file you upload, so that file's content is part of the request. Decide which teams may use each feature on which material.
  2. Where is that content processed? Your Totara site may be hosted in the country you require, but a generative request is processed under your OpenAI account, on OpenAI's terms. Check those terms against your data residency obligations, because hosting location alone does not answer the question.
  3. What about personal information? Totara states that its AI integrations access only the content and data a customer approves, and do not include personally identifiable information. Confirm how that applies to your configuration, and in particular to anything users paste into a prompt.
  4. Who reviews the output? Totara's principles call for AI content to be labelled, logged and verified by a person. Name the role responsible before the feature goes live, not after the first wrong answer.

Where Totara's AI stops

Stated plainly, because it will decide some evaluations:

  • No learner-facing AI tutor or coach ships with Totara. Partners can build one, but it is a project.
  • No automated skills inference. Totara has skills and competency frameworks, but a person maps them.
  • A smaller packaged set than AI-first learning platforms, and a pace of new AI features to match.

If your brief leads with any of those, weigh it honestly. If it leads with control over data, cost and configuration, Totara's approach is a strength rather than a gap.

Frequently asked questions

Does Totara have AI? Yes. Totara documents six AI capabilities: the AI form assistant, standalone AI image generation, AI knowledge check-in, AI course tag recommendations, the AI goal assistant in Totara Perform, and AI-powered recommendations in the catalogue.

Which version of Totara includes AI? The generative features first shipped in Totara 19.1 in July 2025 and are available as AI plugins in Totara 20, the current major version. The underlying OpenAI integration arrived in version 18.8.

Is AI included in a standard Totara installation? No. The generative AI plugins are installed separately by your Totara partner and require an OpenAI API account.

Does Totara work with AI models other than OpenAI? The packaged features are built for OpenAI. Totara states that integrations with other providers can be built as partner customisation.

How much does AI in Totara cost? Totara does not charge per AI action. You pay your model provider for usage on your own account. Custom AI work, such as a bespoke learning assistant, is priced by your partner as a project.

Does the recommendations engine send data outside Totara? According to Totara, no. The recommendations engine processes data within the Totara environment. The generative features are different, because they send requests to OpenAI.

Is AI available on TotaraGov? Not at present, according to Totara. TotaraGov is certified to FedRAMP specifications, which do not currently include the OpenAI integration.

Can we turn AI off after enabling it? Yes. AI features are managed individually from the Manage AI page in site administration, and can be enabled or disabled per feature.

See what Totara's AI does with your own content

Bring a course or a policy document and we will show you each feature working on it, and tell you where it will not help. If Totara is not the right fit, we will say so.

Book a demo: https://www.lambdalearn.io/schedule-a-meeting-lambda-suite

Talk to a specialist: https://www.lambdalearn.io/schedule-a-meeting-lambda-suite

Totara, Totara Learn and Totara Perform are trademarks of Totara Learning Solutions Ltd. OpenAI is a trademark of OpenAI. Lambda Learning Inc. is a Totara Platinum Partner.