The Problem Isn’t Finding AI Tools Anymore
There’s no shortage of AI tools promising to make instructional design faster. Some generate entire courses. Others build videos, presentations, diagrams, assessments, or transcripts. Established authoring platforms keep bolting on AI features of their own, and new products seem to show up weekly.
For instructional designers, the hard part isn’t finding AI tools anymore. It’s figuring out which ones are actually worth your time.
Over 12 weeks, I worked through AI from a different angle each week, looking at tools through the lens of instructional design rather than just testing how impressive their outputs looked. My list included Mindsmith, Claude, Manus, Miro AI, Napkin AI, Otter.ai, Vyond, Articulate Rise 360, iSpring, HeyGen, and Glean, plus the general-purpose AI tools that have quietly become part of my regular workflow.
I kept coming back to one practical question: would I actually use this as an instructional designer? That turned out to be a far better test than asking whether a tool could produce something quickly.
Some tools solved problems I run into constantly. Some produced impressive results that needed enough correction to make the time savings questionable. A few changed how I think about where instructional designers can spend less time producing materials, and more time evaluating, designing, and making decisions.
Here’s what I’d actually keep in my toolkit, where I’d use each tool, and where I’d stay cautious.
Skip the “AI Tech Stack”
There’s a real pull toward collecting AI tools: one for writing, one for presentations, one for video, one for diagrams, one for research, one for courses, one for meetings, one for brainstorming. Before long you’re paying for eight platforms doing overlapping jobs.
I don’t think most instructional designers need that. I’d rather choose tools based on the problems that actually keep showing up in my work. Ask yourself what takes too much time, what you keep building from scratch, where you consistently get stuck, and what AI could help you explore faster without making the design decision for you. Those four questions produce a much shorter, much more useful list than “what’s trending this month.”
1. Claude: My Thinking Partner
The most useful thing I did with AI during these 12 weeks wasn’t asking it to create finished products. It was using it to think through problems.
Claude works especially well for longer documents, weighing alternatives, questioning an approach, organizing messy information, or developing an idea that isn’t fully formed yet. For an instructional designer, that’s useful at nearly every stage of a project: reviewing SME notes for gaps, comparing two approaches to a learning problem, challenging assumptions in a course design, generating scenario variations, checking an outline for alignment, figuring out what to ask a stakeholder, working through a pile of qualitative feedback, or critiquing an assessment strategy.
The difference between generating and thinking with AI matters a lot here. Instead of asking, “Create a leadership course for new managers,” I’d give it the business problem, the audience, the constraints, and my initial thinking, then ask something like: “What assumptions am I making about why these managers are struggling? What would you want to know before recommending training?” That answer is far more useful to me than a finished course outline.
Where I’d be careful: AI can make a weak idea sound polished. A beautifully structured response can create the impression that the reasoning underneath it is stronger than it actually is. I wouldn’t treat Claude’s recommendation as the design decision. I’d use it to help me stress-test the decision instead.
Would I use it professionally? Absolutely.
2. Mindsmith: A Genuinely Interesting AI-First Course Builder
Mindsmith puts AI much closer to the actual course-development workflow. You can use it to build lessons, generate content, develop questions, and move from an idea to something resembling a working learning experience surprisingly fast. That makes it interesting for rapid development. If I were building a short internal training module, a client prototype, or an early version of a course, Mindsmith would earn a spot in the process.
This is also exactly where instructional designers need to slow down. Generating a course isn’t the same as designing one. An AI-generated lesson can have reasonable objectives, logically organized content, knowledge checks, examples, and a clean summary, and still look complete without answering the questions that actually matter: does this training solve the right problem, do learners even need training, are the activities suited to the performance you expect, does the assessment provide meaningful evidence, and are learners practicing the right skill? AI can generate every component of a course before anyone has answered a single one of those.
I’d use Mindsmith for rapid prototyping, not automatic instructional design. Let AI build something fast, then evaluate it hard. That beats staring at a blank screen, as long as the instructional designer stays the one responsible for the design.
Would I use it professionally? Yes, particularly for prototypes and rapid-development projects.
3. Articulate Rise 360 AI: Useful Because It Lives Where You Already Work
One of the more practical shifts in instructional design might not be a brand-new AI platform at all. It might just be AI showing up inside the tools you already use. Rise 360 is a good example. If you’re already building courses in the Articulate environment, its built-in AI cuts down on the friction of drafting and revising, since you’re not generating content in one system, copying it into another, reformatting it, and picking the work back up somewhere else.
That matters more than it sounds like it should. The more AI becomes part of your existing workflow, the less appealing a separate standalone AI tool becomes. Inside Rise, I’d use AI to generate an initial content block, revise an explanation, build question variations, generate examples, tighten up content, brainstorm interactions, or produce a rough first pass I then go in and edit.
Where I’d draw the line: I wouldn’t ask AI to build a course and assume the design is finished just because the result looks polished. The hardest parts of instructional design still happen before and around the authoring tool, not inside it.
Would I use it professionally? Yes, especially if Rise is already part of the project.
4. iSpring AI: Practical for PowerPoint-Based Workflows
iSpring sits in a slightly different lane. It works well for designers and organizations already leaning heavily on PowerPoint-based course development, which matters more than it might sound like on paper. The best AI tool isn’t always the one with the longest feature list. Sometimes it’s the one that fits a workflow your organization can actually maintain long term.
AI built into an existing authoring process can help with drafting, assessments, and content development without asking an entire team to learn a new platform. For organizations producing straightforward compliance, product, procedural, or employee training, that’s genuinely valuable.
My test here is simple: does the AI feature save enough development time to justify using it, while still leaving me in control of the instructional decisions? If yes, I use it.
Would I use it professionally? Yes, particularly in organizations already using iSpring.
5. Vyond AI: Practical for Training Video
Video production eats a surprising amount of instructional design time. Writing the script is only the start. You still need characters, scenes, voiceover, timing, transitions, editing, and revisions. Vyond has long been a solid way to build animated training videos without a full production team, and its AI features speed that process up considerably.
I could see it working well for workplace scenarios, customer-service examples, compliance situations, process walkthroughs, short concept videos, and difficult-conversation demonstrations. Instead of describing five common mistakes managers make during a performance conversation, you can show one happening. That’s often more useful instructionally than another paragraph of text.
Where I’d be careful: video is easy to overproduce, and AI makes that even easier. A three-minute animated video doesn’t automatically make a course better just because it looks more engaging than text on a page. Before I build one, I’d still ask what the video lets learners see or experience that another format can’t. If I can’t answer that, I probably don’t need the video.
Would I use it professionally? Yes.
6. HeyGen: Impressive, but I’d Use It Selectively
HeyGen raises a different kind of question. AI avatars can produce polished presenter-style videos without cameras, studios, actors, or repeated recording sessions, which opens up some genuinely interesting possibilities. For organizations that need content in multiple languages, frequent video updates, or large volumes of presenter-led material, the efficiency gain can be substantial.
I could see real use for multilingual training, short announcements, software or process updates, standardized introductions, and informational videos that get revised often. But I wouldn’t automatically swap out a real instructor, leader, SME, or facilitator for an avatar. Sometimes who delivers the message is part of the message. A welcome video from a program director serves a relational purpose an avatar can’t replicate just by saying the same words in the same order.
The question I’d ask: is the presenter mainly delivering information, or is their presence part of the point? If it’s scalable information delivery, HeyGen becomes a lot more attractive.
Would I use it professionally? Yes, but selectively.
7. Napkin AI: Surprisingly Useful for Visualizing Ideas Fast
Instructional designers need simple visuals long before they need polished graphics. You need to explain a process, a relationship, a framework, a sequence, or a model, and that can turn into an irritating amount of time spent nudging boxes and arrows around a slide.
Napkin AI takes text and turns it into a visual quickly. I wouldn’t expect every generated visual to be ready for a final course, and that’s not really where the value is anyway. The value is in getting from “I need some way to show this relationship” to “here are several visual approaches I can actually react to.” That speeds up storyboarding and early design considerably. I’d use it for process visuals, conceptual relationships, timelines, comparisons, simple frameworks, presentation graphics, and job-aid concepts.
Would I use it professionally? Yes, especially during design and prototyping.
8. Miro AI: Useful When the Work Is Messy
Some instructional design work starts with a tidy project plan. Most of it doesn’t. You end up with interview notes, SME comments, sticky notes, learner feedback, half-formed ideas, requirements, questions, and contradictory stakeholder requests, all in one pile. This is where Miro AI gets interesting. If your team already uses Miro for workshops, mapping, or design sessions, AI can help organize that pile and spot patterns in it.
I’d use it during needs analysis, stakeholder workshops, learner journey mapping, content organization, affinity mapping, brainstorming, and early curriculum design. Again, I wouldn’t let AI’s organization of the information become the final interpretation of it. Grouping 30 comments into five themes is genuinely useful. Deciding what those themes actually mean for the design still takes a person.
Would I use it professionally? Yes, when the project benefits from visual collaboration.
9. Otter.ai: Useful Mostly Because It Removes Administrative Work
Not every valuable AI tool needs to transform instructional design. Some just take work off your plate that you’d rather not spend time on. Otter.ai fits that category. For interviews, SME meetings, discovery calls, or design discussions, transcription is genuinely useful. Instead of splitting your attention between listening and taking detailed notes, you can stay in the conversation and go back to the transcript afterward. For anyone running needs analysis or pulling content from SMEs, that saves real time.
The one caution here is obvious but worth saying anyway: meeting transcripts can contain confidential, personal, or proprietary information. Before I record or upload anything, I’d want to know exactly what my organization permits and how the platform handles that data afterward.
Would I use it professionally? Yes, when privacy and organizational policy allow it.
10. Glean: More Interesting for Organizations Than Individual Designers
Glean caught my attention for a different reason entirely. Instructional designers often assume that when employees can’t find information, the fix is training. Sometimes the real problem is access. The information already exists, employees just can’t find it when they need it. Enterprise search tools like Glean can change that by helping employees find information across an organization’s systems.
That raises a question worth sitting with: does this person need to learn this information, or do they need reliable access to it at the moment they need it? Those are two different problems. If an employee needs to remember a safety procedure during an emergency, training is essential. If they need to look up the current travel reimbursement limit twice a year, better access to information probably beats another course.
Glean is less interesting to me as a course-development tool and more interesting as part of the broader performance-support picture.
Would I use it professionally? In the right organization, yes.
11. Manus: Interesting for Rapid Presentation Development
Presentation development is another spot where AI saves real production time. Manus can move you quickly from raw information to a structured presentation, useful for training decks, stakeholder presentations, project proposals, and early concepts. I’m most interested in it for building a first version. A blank presentation takes time to get moving. A generated one gives you something to react to and tear apart, moving slides, cutting weak sections, rewriting explanations, reordering the sequence, rethinking the visuals. That’s usually easier than starting from nothing.
The danger is accepting the structure just because it looks finished. AI-generated presentations make decisions about hierarchy, emphasis, and narrative on your behalf, and those are design decisions that deserve a second look.
Would I use it professionally? Yes, mainly for first drafts and prototypes.
12. The General-Purpose Tool Still Matters Most
After working through the specialized tools, I landed somewhere I didn’t expect. I don’t need AI to do every part of instructional design for me. The AI I find most useful is the AI that helps me think better about the work, not the AI that hands me the most finished output.
A general-purpose tool can question my assumptions, generate alternatives, test an argument, work through a scenario, organize research, or critique something I’ve already built. Specialized tools have a hard time matching that flexibility. It also puts more responsibility on me, because there’s no “generate good instructional design” button. I still have to know what I’m actually trying to accomplish, and honestly, I think that’s a good thing.
How I’d Build My AI Toolkit Today
If I were starting from scratch, I wouldn’t subscribe to all 12 of these. I’d build around functions instead: one strong general-purpose AI tool for analysis, ideation, critique, and drafting; my existing authoring platform’s AI features, if they genuinely save time; a video tool like Vyond or HeyGen when video is a regular part of production; a visual tool like Napkin AI when diagrams and explanatory graphics come up often; and a transcription tool if interviews and SME meetings are a regular part of the job.
Everything else has to earn its place. If I use something twice a year, I don’t need another subscription for it.
The Test I Use for AI Tools Now
After working through AI at every stage of instructional design, I’ve stopped asking whether a tool is impressive. Plenty of them are. I ask five questions instead.
Does it solve a problem I actually have? A clever feature by itself isn’t enough.
Does it save meaningful time? If I spend 20 minutes generating something and 40 more correcting it, the efficiency argument falls apart.
Can I control the output? I want AI to speed up the work, not push me into its version of the approach.
What am I giving up? Saving time by taking over a decision I should be making is a very different thing from saving time on repetitive production work.
Would I pay for it? This might be the most honest test of the five. A tool I enjoy experimenting with isn’t automatically a tool I need professionally.
What Actually Changed After Testing These Tools
I went into this expecting efficiency to be the headline. It wasn’t. AI absolutely makes instructional design faster, and it can dramatically cut the time it takes to produce drafts, videos, graphics, scenarios, and assessments. But speed turned out to matter less to me than where the instructional designer’s judgment still has to show up.
The more AI can produce, the more it matters that someone decides what’s actually worth producing. Someone still has to determine whether training is even the right solution. Someone has to identify what learners need to practice. Someone has to decide whether an assessment measures anything real. Someone has to catch it when a polished AI-generated explanation is wrong, incomplete, biased, or just unnecessary.
That’s where I see the instructional designer’s role getting more important, not less.
I’ll keep using AI. I’ll use it to get past a blank page, build alternatives, speed up production, organize information, put together prototypes, and question my own ideas. But I don’t want AI making every instructional decision just because it’s capable of it. The most useful AI tool isn’t the one that produces the most. It’s the one that leaves me more time and attention for the parts of instructional design that still require me to actually think.
Let’s Talk About Your Program
If you’re trying to figure out which AI tools actually belong in your organization’s instructional design workflow, and which ones are just adding noise, that’s worth a closer look before you commit budget to another subscription.
I review and analyze existing programs and courses, then provide clear recommendations for strengthening and modernizing them, including where AI tools genuinely earn their place and where they don’t.
Get in touch through the contact form on my website at https://yourelearningworld.com/contact/ and let’s take a look at what your program needs.
Leave a Reply