With the increase usage of AI in learning systems, whether they are an LMS, LXP, Learning Platform, a combo of LMS/LXP, Talent Development system, Mentoring platform, Coaching platform, or Sales Enablement Platform, where the whole focus is on sales training, or some other learning system that doesn’t want to call themselves any of the above and instead comes up with something they think is clever, but nobody searches nor asks for it.

As of today (late November 24), the industry is witnessing an evolution of AI trends in learning systems.

It’s important to note that AI, despite its potential, is still in its early stages.

Nobody knows how fast it will change in 2025, what will benefit industries, or what will scare the bejeebers out of people, including those who think AI is doom.

I looked solely at our industry and learning systems.

Not learning tech products, because the market isn’t at a mature stage, plus the usage levels, depend on the offering, updating, and approach.

You can be hot today and be a dumpster fire tomorrow. I’ve seen a lot of dumpster fires – even ones where raised capital would make your mouth drop.

EdTech—K-12 and higher education—is an audience, but the type of learning system is the same.

In other words, if Bonzo University is using an LMS, by vendor Dog, then you, Corporate could be using the same system.

The Vision 2025

The best way to explain what I see as the Vision 2025 for learning systems, to view it – and then break down all the points within it.

You can download my wonderful PDF – as a souvenir or you want to read and decipher with this post. I didn’t add a watermark, because I’m not a huge fan of them.

I did though add my wonderful name on it, because you will be able in a year from now go “That’s Craig’s Vision, and he was right. Where is my dinner?”

The Entire Process

The Breakdown of AI Levels

The beginning is the term “learning systems” which is an umbrella term for all the different types of learning systems – such as an LMS and so forth. In this first part – the umbrella appears followed by one ofThe three for the Vision of 2025 are based on trends in the market and where I see vendors will slide under, showcasing the promising potential of AI-driven learning systems in the near future.

AI-Driven, AI Options – with some splitting into AI Focus, Little to no AI functionality

AI-Driven

This term will, I believe, identify those learning systems, regardless of type, that present themselves as AI-Driven LMS, AI-Driven LXP, AI-Driven Talent Development, and so on.

What separates them from the rest of the pack is not only their commitment to AI but, more importantly, the level of AI functionality they are adding to the system.  

There are quite a few out there for 2024, and if you can see their roadmaps or talk to them, it is obvious that they are pushing ahead, recognizing that the future of learning, online learning here, will require a boost with AI to achieve the next level of training and learning—regardless of whether it is customers/clients, employees, students, association members, and so on.

Who Stands Out Early On – AI-Driven

My Top Four – #1 is below

• Learnster – Impressive items with a lot coming out by the end of this year and early next year. The document report with AI is cool. They have the ability for a Q/A – which will be, for many vendors in our space, the “next big thing” to do. Again, there are vendors who have it – some are better, others it is just okay, next. I like Learnster’s approach.

I like the library each Learner can use, where they can add their own AI media to the system.

They are developing personal agents, which go beyond the co-pilot stuff you hear about, and is underwhelming – but comes off as “cool.”  

A personal agent helps the Learner by recognizing what items it can do and allow the Learner to focus on other.

A personal agent can be a coach or mentor to the Learner as the Learner uses the system. When you think of personal agents (autonomous agents – the goal you want), you have to think about it from the learning standpoint – for the Learner.

Learning needs to be about the Learner and what they can do in the system, not the person overseeing e-learning or the admin.

I digress.

Big Early Wins – And this is just a few of them

• Learnster AI Learning Companion – In basic terms, it taps into the data of the Learner as they are using the system (yet still protection and you – admin/overseer, identify what data cannot be seen or utilized – I recommend talking to them to know more)
• Learning Spaces: AI identifies the topics and subjects to explore based on the learner selections themselves—hard to explain, cool to see.
• The learner, based on the initial response of the AI, can decide what they want to do next – choose a quiz, create a presentation, create an online document, and much more. You can even do something they call a “remix.
• AI podcasts with a synthetic voice are coming soon. They will be tied around the topics the Learner is interested in. You can even remix them.
• The entire AI approach is 100% based on informal learning, which so many systems say they are but, surprise, are not.
• AI translation is in the works, too.

This is AI-Driven – The Other Three

• Fuse – While I cannot identify what they have in the works with AI, I came away very impressed in a few areas and intrigued by others.
• IMC-AG – A multi-agent (persona, in a way, beyond co-pilot) – expected to roll out in in Q2, 2025. If interested, chat with them. And again, other items are in the works. I was surprised by how far they are with AI compared to the market.
• Cypher learning – Their authoring tool – the advanced one, an ID or someone with intermediate skills that utilize this. They also have a co-pilot (legit and different than personal agents) and AI assistant (which is intriguing, but they never tell people that it may output fake or false information and you should verify things – which is surprising since they note the warning in numerous other areas), personal agents – coming soon. This is heavily focused on the Learner – as it should be.

AI Options

AI Focus are vendors who list the usual I see all the time (okay, nowadays) and then have others in the works or have a couple that vendors do not have (in general, skills is a growth AI area—and yes, the top two are right in the mix), but I can see at least one of these vendors saying AI-driven.  

And no, I won’t mention them.

Here are vendors that I surmise will push AI Driven, that right now, what I see and what I am aware of, slides them into my next category – AI Options.  

This is really a split between AI Focus and then the options side.

AI Options – I see a few as AF (which can they jump into AI driven?

Any system can, but it needs more than five feature sets.  

Those that are AI-driven—wow, there is a big difference between them and AF and AI options.  

Regardless, most of the market is here – either AF or AO. This is just a short list – so don’t disappear if you are not here.  

There is one that has AI in it, but just because they have raised a lot of capital means little to me.  

I overwhelmingly believe it is going into their AI solution and not per se into the learn solution.

Thought Industries

Docebo

Cornerstone Learn

KREDO learning platform

SanaAI, the Learn system—surprising because they have AI in their name, but they have two offerings: one is the Learn side, and the other is this AI Assistant capability—which is, well, AI Assistant, a total separate entity.

At this time, you cannot integrate the AI Assistant (nor can your learn salesperson help you with it – seeing it, etc. – yep, they have another salesperson for that – worst idea ever) – that they are continuing to boost up with capabilities, and the learning solution – which has AI offerings – but a chunk is around their content creator. They do have an AI Q/A – which is nice, using generative AI – again, this, I believe, will be the “next” feature vendors in the AI Options will add.

Absorb Learn

LearnUpon

The Splinter

When you commit to AI, whether you are AI Driven or AI Focus – the Overhaul of UI/UX is a must.

There is no, well, we can stay with what we have and just roll with it. No, No.

It’s pretty simple- you either do it, or you don’t.

I believe those that do, well again push ahead.

I understand that cost is always involved here, which is why I would focus first on the learner side, and then on the admin side (which is what vendors do anyway).

There is no doubt in my mind, that AI with metrics is coming at a level that hasn’t been seen before, because of AI’s capabilities here, and where it will go, in the next year and year thereafter.

Already one system I have seen, has some AI in their metrics. It’s cool.

And the potential is massive.

Thus, on the admin side, on the metrics side, it has to be tweaked to deliver. Nobody likes wow, and on the back end, blah.

When you get to UI/UX, the idea that you shove as much as you can on the front is a huge mistake.

You can easily streamline (which is what I would do), then on another screen have the stuff you want to showcase the learner to see – which is the content and so forth.

That said, as an industry, the long in the tooth, Grid look is getting stale.

New UI/UX Example

Fuse’s next UI/UX coming out in 2025.

It is streamlined.

The top has the Q/A AI gen item. Which makes 100% sense. Get that front and center. Let the learner ask questions, see retorts, narrow down see retorts.

Perhaps, the retort goes somewhere into the system – a path, some content or skill or a video or 3rd party content with that specific subject matter or whatever.

Always remember that just because a vendor trains the LLM on their own data, and you then place your content into the LLM, it may still make mistakes.

I have listened to vendors make the argument that isn’t true. Yes, yes it is.

The inherent flaw of AI are hallucinations.

Even AI bias is a problem.

No LLM on the planet, not home made, commercial or open source, trained on whatever you choose – may generate hallucinations.

Even if a vendor uses a RAG, pitches guardrails, has multiple LLMs (which I always recommend), hallucinations exist.

Oh, and that LLM agnostic thing some vendors are now saying?

100% not true in 2024.

That said, a Model As A Service, I believe is the way to go. Okay, let’s move on.

Could a vendor limited to no options go straight into UI/UX overhaul in 2025, with the approach that they plan to get to the next step or the higher step?

Absolutely.

That said, trust me, a lot won’t do it. Ditto on the ones who are AI Options, even AI Focus. Would I want to buy a system that has zero plans to update their UI/UX in 2025? No.

What if they updated in 2024? They will still need to make some tweaks for 2025, due to AI.

This isn’t a skirt and hide, and tada!

Business Productivity

This does not mean with the system you boost productivity – because every system can do this, without AI.

This is saying that within the learning system (I will use an LMS just as an example; it can be any type of learning system)

• Today, companies that have implemented AI to their employees have, overall, no idea how often their employees are using AI – especially if the employee continues to use ChatGPT or any other that is free – and there are a lot out there – thus you may block ChatGPT at your company, but how do you block others? Plus, there is no way you can block it at their home or on a mobile device – that isn’t from the company.

Do not underestimate how an employee, who is a human being, can access various LLMs – AI here to use to build a report, complete documents, send e-mails (with a copy and paste), create and send over presentations, and so much more.

A lot of people do this at home.  

Even if you block ChatGPT and tell them you can’t use another, unless you block them on the entire internet, they still—okay, some will—go elsewhere and do a cut-and-paste.

Heck, you may get prompt leaking within your own AI you have added with whatever LLM. A new can of worms.

Okay, back to productivity.

We know that using a Talent Dev system, HRIS is at the forefront, even with any system out there, including LMS and Mentoring.

What BP says is okay; you are connecting with various business tools (not HR-related, nor Business, but productivity-related).  

You connect with Office365. You connect with Google Workspace.  

You connect with OneDrive, Google Drive, Google Worksheets, etc.

Employee access is with Office365, which includes PowerPoint, Excel, and Word.

How do you track what is being utilized with Gen AI?

You don’t unless you embed an AI tracker within each product.

Each of those offerings has AI options, and I am not referring to the use of Co-Pilot.

If the employee is in a learning system where you have connected Office365 so that they can use it within the system, now you can track when they use the AI in the learning system because the vendor would need to use it, for example, the ability for the learner to create documents, presentations, and so on.

In the Learning System Vision of 2025 – the learner starts to have control in their learning – not the usual approach of Admin dominates the learning angle.

I see vendors incorporating the ability for folks to create presentations and documents (similar to what Learnster is doing) and then publish.

People want that control.  

The AI can track.

Want an employee to do some reflection on the content? Track.

Want the employee to create their own content?

Thus, they get access to the content creator tool, which is 100% a great idea. I have seen this already sans AI—track it.

The spaces angle, let’s say it is a place where they house the content of what they created. Then, they can add it to wherever is doable in the system—reflect, review, refresh—a system with AI tracking can do it.

Heck, Adobe has an AI solution in its system, and there are vendors who can have an employee sign, read, and download a document—it is embedded.  

Then there are those who require a download – and thus, the employee can do similar.  

I like the former.

BP pushes the whole streamline tasks by using the learning system – okay, leveraging it, with the vendor’s multi-agents and personal agents (autonomous), which the vendor wants anyway.

The Vision says a vendor pushing AI may take that step rather than just angling around the HR side.

Already some companies want the learning system vendor to connect to the company’s LLM.

Why not go the extra mile?

I should add that you can be AI-driven and not incorporate all of BP. Heck, none is fine, too.

Then there are vendors that, in reality, are already there, without them realizing it.

The biggest challenge for any entity, even edtech, is getting the person to stay in the system and do other things.

AI changes that, with a Q&A piece on the learner’s side and other options the learner can choose from.

This shouldn’t be limited to formal learning because when you are AI-driven or AI-focused, the learner is the main being here—truly informal learning.

That will get them to stay.

Oh, and those folks all about gamification—AI will tap into it with a vendor utilizing it. I think AI in 2024, with gamification as a vendor, would use it, and some vendors use machine learning, which is a form of AI.

Vision 2025, says if you are no on overhaul of UI/UX then the Vision of 2025 ends for you. If you decline BP, vision ends. If you decline both Vision 2025, ends.

To Deliver the Vision of 2025 – five core items that are needed.

No exceptions.

And all of them make 100% sense.

• Reflective – this shouldn’t just be in a piece of content. It should be in numerous areas to stimulate thinking. Reflective is deep thinking. Right now, AI can’t do deep thinking.

If a learner is listening to a podcast, or watching a video, or discussing ideas in a cohort, or using an AI Coach (another hot item in 2025) or an AI mentor, add reflective here.

Never limit deep thinking. It stimulates a thought process, helping with retention.

• AI Q/A – with the extension of multimodal – There are quite a few LLMs out there that are multimodal. Just text generative is no longer the king. I totally understand why vendors are not at multimodal, or just using PDF for summaries (a couple of vendors do this today – it either looks slick or not). AI Q/A with text is effective, especially when the learner can do a deeper dive and it can go to specific areas in the system.

Multimodal pushes the envelope far more. The premise I type in some text and it outputs a video (they are not long, and the time to view it, is not quick), would be of use in some entities, but not all. The tech around this from an AI standpoint will evolve.

I know folks that are using 3rd party platforms integrated to get better capabilities and multimodal. However, there are some LLMs out there, that can hit the multimodal quite well, especially when they open it up, so that people can create capabilities – I can’t recall what they refer these people as – that can be free or fee-based.

When you recognize that Open Source LLMs will catch up to Commercial, depending on the vendor, and even today, there are some that are really impressive (and vice versa), a community can develop and you can pick what you want in that LLM.

The cost? Probably nothing to use it.

Lastly you want to use a 3rd party you are confident will be around in say three to five years. This industry is already flooded with AI offerings, and just like dot com days, there are ones getting a lot of capital, and still failing.

• UI/UX – Already discussed.
• Personal Agents (autonomous) – discussed in other posts – here and here.
• Revamped Metrics – Ditto
• Other AI items not listed? – The AI world has no limits – okay quality, use case to have it in there, benefit to the learner specifically (remember informal), and token costs – will tamper that down. Still, just the idea of what is coming, what is here – a vendor hasn’t used it, or what they are doing – is open.

Bottom Line

There is a lot to digest here. A lot.

However, the Learning System Vision 2025

Never was meant to be short and sweet.

If you wanted that, here is a summary,

“The blog post titled “Vision 2025 by Craig Weiss” outlines a futuristic perspective on learning systems and their integration with artificial intelligence (AI).

• AI Integration: The document emphasizes that many learning systems will possess some AI capabilities, but not all will be AI-driven. It suggests that as AI technology advances, more learning systems will integrate AI functionalities necessary for achieving the outlined Vision 2025 goals.
• AI Capabilities and Focus: Some vendors are expected not to prioritize AI due to market trends, which will result in systems with minimal to no AI functionalities. In contrast, others will heavily focus on AI, enhancing user interfaces and user experiences (UI/UX) to be learner-centric with streamlined, AI-driven operations.
• Enhancements and Goals for 2025: There will be an overhaul in metrics and a revamp of deep thinking strategies, with additional learning and training metrics linked to AI.
• Features and Learning System Types: The document discusses the inclusion of multimodal options, personal autonomous agents, and a variety of learning system types such as LMS (Learning Management System), LXP (Learning Experience Platform), coaching, sales enablement, and more, each focusing on core aspects of training and learning.

This vision reflects a shift towards more AI-driven functionalities in learning systems by 2025, with a broader and more integrated approach to online learning aka e-learning.

(And like I said, always check to make sure of accuracy)

Yeah.

I did that.

E-Learning 24/7