Totara Thursdays Rapid Fire Edition: AI in Totara – Purposeful, Governed, and Useful

Rapid Fire articles show up from time to time inside the Totara Thursdays series. The approach is simple. Take one practical domain and break it into a few focused questions. Each one gets a clear answer, then a short reflection on what it actually means for how a learning system works. No heavy theory. Just enough thinking to see how these choices play out in real organisations. In this post, we break down Totara’s approach to AI in the LMS.

Across the platform, Totara’s approach to AI is deliberately practical. AI can help learning teams write course content, generate supporting images, organise resources with automated tagging, surface relevant learning through personalised recommendations, turn existing content into knowledge check-ins, and help managers structure clearer development goals. These are practical capabilities that remove friction from everyday learning and performance work, while keeping organisations in control of how, where, and by whom AI is used. AI functionality can be enabled selectively, managed through permissions, connected to approved language models, and rolled out at a pace that aligns with internal governance and operational readiness.

On the surface, the opportunity can look straightforward. Less manual effort. Faster content creation. Smarter discovery. Better structured conversations.

It is easy to assume the real question is simply which platform has the most AI features. But in enterprise learning, that is rarely where the real story begins.

The moment these capabilities start interacting with live learning environments, existing content structures, development frameworks, and manager-led workflows, something deeper starts to happen. Activity increases very quickly, and that can create the impression that learning capability is improving just as quickly.

But activity is not the same as capability, and capability is not the same as performance.

This creates a different question altogether. Not just what AI can do, but what happens when practical AI capabilities meet real organisational complexity.

Why is that only the starting point?

One of the first things organisations notice about Totara’s AI capabilities is how quickly friction starts to disappear.

Course descriptions can be drafted in seconds. Learning objectives become easier to refine. Images can be generated inside the platform. Course tags appear automatically. Relevant learning starts surfacing with less manual effort. Managers can use guided prompts to structure clearer development goals.

That is genuinely useful, and in most organisations those gains become visible fairly quickly. Content teams move faster, administrative effort starts to drop, and the platform begins to feel more responsive.

But speed creates its own illusion.

When more content is created, more resources become discoverable, and more learning activity starts showing up in the platform, it can feel like capability is improving.

However, activity is not the same as application.

Learning only builds capability when the right content is applied in the right context, reinforced through reflection and conversation, and connected to real work over time. Without that connection, AI may reduce friction, but it does not automatically improve performance.

In practice, teams can create twice as much content in half the time, while learner completion, manager follow-through, and on-the-job application remain largely unchanged.

So AI can accelerate learning activity very quickly. Capability still has to be built the hard way.

What’s actually happening beneath these AI capabilities?

This is where the conversation starts to shift.

Capabilities like Totara’s AI Writing Assistant, SMART Goal Assistant, knowledge check-ins, automated tagging, and personalised recommendations can look like productivity tools. They help teams move faster, reduce manual effort, and make useful actions easier to repeat.

Useful as that is, it is only part of the story. These capabilities are not replacing learning design, performance management, or organisational decision-making. They are interacting with systems that already exist, and in doing so, they start revealing how well those systems were designed in the first place.

Here’s why.

If content standards are inconsistent, automated tagging quickly exposes weak metadata. If development frameworks are unclear, even well-written AI-assisted goals struggle to translate into meaningful performance conversations. And if managers are not already coaching consistently, smarter prompts do not suddenly create better follow-through.

In practice, AI-generated goals can look strong on day one because they are clear, measurable, and aligned to development priorities.

But six months later, no check-ins have happened, no coaching conversations have been recorded, and the original goal has quietly become another completed activity.

So the platform is not just becoming smarter. Its existing strengths and weaknesses are becoming harder to hide.

Why does that matter more in enterprise learning than in consumer platforms?

AI recommendations feel familiar because most of us already experience them every day.

Streaming platforms suggest what to watch next. Online retailers recommend what to buy. Social platforms surface what is most likely to keep our attention. In those environments, success is usually measured through clicks, views, engagement, or time spent.

That model works well in consumer platforms, where attention is often the primary metric.

Enterprise learning operates under very different conditions, and Totara’s architecture reflects that. AI capabilities can be enabled selectively, controlled through role-based permissions, audited over time, and deployed within customer-approved AI environments, giving organisations control over adoption without compromising data boundaries or governance requirements.

A recommended course cannot simply be interesting or relevant. It may need to support a compliance requirement, reinforce a critical capability, align to a role-based development path, or feed into a manager-led performance conversation. In many cases, it also needs to remain visible, traceable, and auditable long after the recommendation was made.

This creates a different standard for AI. Relevance still matters. But relevance on its own is not enough when learning is connected to risk, regulation, operational capability, and business priorities.

In practice, an AI recommendation may surface a highly relevant course based on learner behaviour. But if that recommendation bypasses a required certification, ignores audience rules, or disrupts a structured development pathway, it creates friction instead of progress.

So in enterprise learning, AI is not just shaping engagement. It is shaping accountability.

What does AI actually reveal inside an organisation?

The short answer is fairly simple. AI rarely creates new organisational problems. What it does is make existing ones much harder to ignore.

The moment AI starts generating course content, recommending learning, tagging resources, or shaping development conversations, operational gaps begin to surface much faster than they did before. Issues that once sat quietly in the background suddenly become visible because AI depends on structure, consistency, and clear rules to work well.

This creates a very practical maturity test.

Duplicate content becomes easier to spot because recommendations surface multiple versions of the same programme. Weak metadata becomes obvious when search results feel inconsistent or irrelevant. Ownership gaps start to matter when no one is quite sure who should approve AI-generated content or review automated changes. Permissions become critical when some users can generate content, while others should only consume it. And as usage grows, token consumption, API costs, and access control become operational decisions, not technical details.

In practice, we often see AI recommendations underperform for a very simple reason. The content was already fragmented, poorly tagged, or inconsistently maintained long before AI was introduced.

So when organisations say AI is not performing as expected, the issue is often not the model.

It is the maturity of the environment the model has been introduced into.

That is why AI does not create organisational maturity. It exposes it.

What should organisations really be evaluating when looking at AI in Totara?

This is usually where platform conversations become too narrow.

Teams compare feature lists. One platform offers AI writing. Another offers recommendations. Another adds image generation, tagging, or goal support. The conversation becomes about capability coverage, feature depth, or who appears to be moving faster.

That feels like the right place to start. In practice, it is rarely the right place to stop.

The better question is not, “What AI features does the platform have?”

The better question is, “Do we have the structure, permissions, content standards, approved AI environments, and performance discipline to use capabilities like AI writing, automated tagging, personalised recommendations, and goal support at scale?”

A writing assistant is only useful if content ownership is clear. Automated tagging only improves discovery if metadata standards already exist. AI-generated goals only strengthen performance if managers are expected to review, coach, and follow through. And recommendations only become meaningful when learning pathways, audience rules, and compliance requirements are already well defined.

In practice, mature organisations often start small. They may enable AI only for specific user groups, connect Totara to an approved language model, monitor usage through controlled access, and expand capability only once governance and confidence are in place. AI content generation may be enabled only for trained content owners or site administrators before access is expanded across broader teams.

So the real evaluation is not platform capability alone. It is operational readiness.

What does that enable in practice?

Once the foundations are in place, AI stops feeling experimental and starts becoming operational.

When permissions are clearly defined, content standards are understood, ownership is visible, and managers are already accountable for development conversations, capabilities like Totara’s AI writing, goal support, knowledge check-ins, automated tagging, and personalised recommendations become highly practical.

Course creation becomes faster because content owners are working from consistent standards. Goal setting becomes stronger because managers are building on existing coaching habits, not trying to create them from scratch. Knowledge reinforcement becomes easier because learning resources can be turned into low-friction check-ins that support reflection over time. Discovery becomes more relevant because the underlying metadata is already structured and trusted.

The real gain goes beyond efficiency. It creates scale without losing control.

In practice, we often see organisations start with a small pilot group of content creators, site administrators, and line managers. They test real use cases, monitor adoption, manage usage costs, and refine access rules before expanding AI more broadly across the organisation.

So the strongest AI adoption rarely starts big. It starts disciplined.

What does Totara’s approach to AI actually tell us?

By this point, the pattern is usually hard to miss.

Totara’s AI capabilities are not designed to replace learning design, management judgment, or performance conversations. They are designed to strengthen the systems that already exist, remove friction where work is repetitive, and make good learning and development practices easier to scale.

That is where the real insight sits. AI does not create organisational maturity. It exposes it.

The moment AI starts writing content, recommending learning, shaping goals, or organising knowledge, every strength and every weakness in the surrounding environment becomes more visible.

So the real question is no longer, “Does our platform have AI?” It becomes, “Are we mature enough to use it in ways that strengthen capability, improve performance, and scale with confidence?”

If you’d like to explore what Totara’s AI capabilities could look like in your own environment, let’s talk. Book a short 20 to 30 minute conversation with Lateral, and we’ll help you assess what can work for your organisation, where it can add value, and how to approach it with confidence.