The Best AI Tools for Instructional Design (2026): What I Actually Use

If you want the short answer: I use Claude (specifically, Claude Code) for analysis, writing, and development, the Articulate AI Assistant for authoring inside Storyline and Rise, Magnific for images and media, and devlin.ai for practice conversations. Full disclosure on that last one: devlin.ai is my product, and I held it to the same test as everything else on this list.

This is not a directory of 70 tools I found on other people's lists. Every entry here is organized by the job you need done, and every entry earned its spot the same way: I started a trial and tried to produce real work with it. If you want the bigger picture first, start with how instructional designers are using AI in 2026.

The stakes are real. According to ATD's AI in Instructional Design: Transforming Workflows and Content Creation report (2025), 80 percent of instructional designers already use AI tools while designing courses, part of a broader shift I cover in employee training statistics and trends for 2026. Here's how adoption actually breaks down:

Finding	Number	Source
Instructional designers who use AI tools	80%	ATD, AI in Instructional Design (2025)
AI users who use generative AI specifically	96%	ATD, AI in Instructional Design (2025)
Say AI improves the quality of their course design	70%	ATD, AI in Instructional Design (2025)
Say AI greatly reduced course design time	37%	ATD, AI in Instructional Design (2025)
Don't use AI at all, mostly citing lack of trust in AI output	20%	ATD, AI in Instructional Design (2025)
Instructional designers who have used ChatGPT	84%	Dr. Philippa Hardman's survey research

One more thing before the list. You need 3 or 4 tools, not 50. Most of the anxiety around AI tooling comes from trying to keep up with everything instead of picking one tool per job and getting good at it.

How did I test these AI tools?

Every tool on this list went through the same 10-minute test: start a trial and try to produce work-quality material. If a tool makes slop that takes longer to fix than building the thing by hand, it's out.

1. Start a real trial

   No demo videos or sales calls. I sign up for the free trial or free tier and get my hands on the actual product. If I can't touch it, I can't rank it.

2. Build something I'd actually ship

   I pull a real task from a current project (a storyboard, a course image, a practice scenario) and ask the tool to produce it. I use real inputs and constraints instead of the vendor's canned example.

3. Compare edit time to build time

   Then the only question that matters: does fixing the AI's output take less time than making it myself? If yes, the tool stays. If no, it goes, no matter how good the demo looked.

On qualifications: I don't have an AI or machine learning degree. What I do have is nearly a decade in instructional design, daily AI use, a habit of trying new tools as they come out, and a lot of conversations with instructional designers plus surveys of hiring managers about how their teams use AI. That's the lens for every verdict below.

And to repeat the disclosure: devlin.ai is mine. It went through the same 10-minute test, and I'll tell you where it falls down just like every other entry.

What are the best AI tools for instructional design?

For most instructional designers, the short list is: Claude for thinking and writing, the Articulate AI Assistant for authoring, Magnific for images, ElevenLabs for voice, devlin.ai for practice simulations, Claude Design for building interactive content without the technical friction, and Claude Code once you're ready to move into power user territory. Here's the full field:

Tool	Job to be done	Best for	Biggest limitation	Pricing (checked July 2026)
Claude	Analysis, writing, storyboarding	One assistant across the whole design process	Fewer plugins and integrations than ChatGPT's ecosystem	Free tier; Pro from $20/mo; Max plans for heavy use
ChatGPT	Analysis, writing, custom GPTs	The field default; fast drafts and broad ecosystem	Output plateaus without strong prompting	Free tier; Plus from $20/mo
Microsoft Copilot / Google Gemini	Analysis inside restricted orgs	IDs limited to IT-approved tools	Capability depends on your org's license tier	Usually covered by Microsoft 365 or Google Workspace licensing
Articulate AI Assistant	Rapid prototyping & generation	Teams already living in Storyline and Rise	Image quality trails dedicated generators	Included with Articulate 360 subscriptions
iSpring Suite (AI features)	PowerPoint-based authoring	PPT-heavy teams converting decks to courses	Tied to the iSpring ecosystem	Paid annual license per author
Magnific (formerly Freepik)	AI images and media	Course imagery and upscaling that looks professional	Another subscription to pay for	Paid subscription tiers
Synthesia	Avatar video	Talking-head training video at scale	Avatar delivery flattens emotional content	Free plan; paid plans by video minutes
HeyGen	Avatar video	Quick avatar video and translation	Same category caveat: wrong tool for high-stakes emotional topics	Free plan; paid plans
ElevenLabs	AI voiceover	Natural narration without a recording booth	Costs scale with audio volume	Free tier; paid plans by usage
devlin.ai	Practice and role-play simulations	A scored practice layer inside Storyline, Rise, or any LMS	Launched in 2026, the youngest tool here	Starter $9/mo ($6/mo annually); usage-based credits
Custom GPTs / Gemini Gems	DIY practice bots	Free experimentation with role-play prompts	Lives outside your course; can't score back to the LMS	Custom GPTs need ChatGPT Plus; Gems included with Gemini
Claude Design	Interactive prototypes and visual builds	The fastest on-ramp to AI-built learning experiences	Front-end output only; power users will outgrow it	Included with paid Claude plans (Pro and up)
Claude Code	Custom eLearning and automation	Designers who want to build at scale	Terminal-based; real learning curve	Runs on paid Claude plans or API usage
Zapier / n8n	AI-powered automations	Connecting AI to your L&D stack	Someone has to maintain the workflows	Free tiers; paid plans

Prices and plan names change fast. Treat these as entry points and confirm on each vendor's site before you budget.

Not sure which rows apply to you? Answer a few questions and get a starter stack for your situation:

Which AI tools are best for analysis and instructional design work?

A general-purpose frontier assistant beats every specialized "ID analysis" tool I've tried for needs analysis, learning objectives, and design thinking. The specialized wrappers mostly repackage the same models with a worse interface. For a full walkthrough of where AI fits in each design phase, see using AI across the design process.

Claude

Claude is my primary work assistant. As of mid-2026, I pay for the Claude Max plan and use the latest Claude model (Fable 5 as I write this, and I upgrade whenever a more intelligent model ships) for everything work-related. For analysis work, it's excellent at digesting source documents, pressure-testing learning objectives, and drafting stakeholder-ready summaries. Where it falls down: ChatGPT still has the bigger plugin and integration ecosystem, so if your workflow depends on third-party connectors, you may feel the gap.

ChatGPT

ChatGPT is still the field's default. Dr. Philippa Hardman's survey research found 84 percent of instructional designers have used it, and for good reason: it's fast, capable, and everyone already knows the interface. I still pay for it, though these days I use it mostly for personal stuff rather than client-quality work. Custom GPTs are the underrated feature for IDs: you can build a reusable assistant loaded with your style guide and process docs. Where it falls down: without deliberate prompting, its output drifts toward generic training-speak that fails the 10-minute test.

Microsoft Copilot and Google Gemini

If your organization restricts AI tools, Copilot or Gemini is probably what you're allowed to use, and that's fine. When I've asked my audience whether AI tools are approved, restricted, or under review at their orgs, "restricted" comes up constantly, and no tool list I've seen addresses it. The practical answer: an approved enterprise assistant you can actually use beats a frontier model you'd have to sneak past IT. Both handle analysis and drafting well, Gemini Gems give you the same DIY-assistant pattern as Custom GPTs, and your data stays inside your org's tenant, which is the whole point.

Which AI tools are best for storyboarding and course authoring?

If you author in Articulate 360 and just need to get something online fast, the built-in AI Assistant is the shortest path from draft to working course. If you need higher-quality output, then you should draft in a frontier assistant and bring it all together in your authoring tool of choice (or skip the authoring tool and ship your HTML output as a SCORM package). For the non-AI side of the toolbox, here's the full instructional design software roundup.

Articulate AI Assistant

The AI Assistant generates course drafts, images, text-to-speech narration, and sound effects right inside Storyline and Rise, and it's included with an Articulate 360 subscription. That "inside your authoring tool" is more helpful than any single feature. I use it for quick sound effects, images, and AI voices when I'm moving fast. Where it falls down: the image quality isn't as good as Magnific's, but it works fine for quick prototypes, and skipping a second subscription is a real advantage if budget is a concern.

iSpring Suite AI features

iSpring's AI features live inside its PowerPoint-based authoring suite, which makes it the natural pick for teams whose source material is already decks. Where it falls down: you're committing to the iSpring ecosystem, and if your team's standard is Storyline, the AI features alone aren't a reason to switch.

One note on the "AI will kill authoring tools" chatter: Storyline isn't disappearing anytime soon, and it's still a valuable skill to learn. There's a real trend of practitioners using Claude (and now Claude Design) to generate standalone HTML interactions instead of authoring them traditionally, and it can work. But somebody has to maintain that code later. More on that in the custom development section below.

Which AI tools are best for eLearning media (images, video, voice)?

One pick per medium: Magnific for images, Synthesia or HeyGen for avatar video, ElevenLabs for voice, and the Articulate AI Assistant's built-in media when "good enough with zero extra subscriptions" is the right call.

Magnific (formerly Freepik)

Magnific is what I pay for when course imagery needs to look professional. The generation and upscaling quality is the best I've used for eLearning work, and it replaced a lot of stock-photo hunting for me. Where it falls down: it's one more subscription, and if the Articulate AI Assistant's images already clear your quality bar, you don't need it. Note the rebrand: this is the platform formerly known as Freepik, so old bookmarks and old pricing pages will mislead you.

Synthesia and HeyGen

Avatar video tools shine for talking-head content at scale: policy updates, product walkthroughs, anything you'd otherwise re-record every quarter. Synthesia is the category standard; HeyGen is strong on translation and lip-sync. Where they fall down: avatar delivery flattens emotional content. If the subject is layoffs, safety incidents, or difficult feedback, a synthetic presenter undercuts the message, and you should use a human or skip video entirely.

ElevenLabs

ElevenLabs is my pick for AI narration that doesn't sound like a GPS. For most eLearning narration it has crossed the "learners don't notice" threshold. Where it falls down: costs scale with audio volume, so a narration-heavy curriculum needs a real budget line, and you should still script for the ear (short sentences, no wall-of-text narration).

What's the best AI tool for practice and role-play simulations?

This is the one category where AI does something that simply wasn't possible before: open conversation practice instead of pre-canned branching. The traditional approach to simulations in Storyline is pre-written answer choices and predefined branches, which means planning every branch and scripting every response. When AI powers the simulation, the learner responds naturally and the AI responds dynamically in character.

When the build time for an immersive simulation collapses, practice stops being the thing we cut from scope.

The biggest simulation tool lists in this space skip the L&D category entirely, which tells you how new it is. If you're evaluating options, I've written a deeper guide on how to evaluate AI simulation tools and a comparison of AI training simulations against other practice formats.

devlin.ai (full disclosure: this one is mine)

devlin.ai is my product, so weigh this entry accordingly. Some history explains why it exists: I had tried just about every AI tool aimed at instructional designers, and the tool I wanted didn't exist. Later I built Custom GPTs for my own program, and they worked as role-play coaches, but they lived outside the course and couldn't report anything back. devlin.ai was built to be, essentially, an embeddable version of a custom GPT, made specifically for instructional designers, that specializes in role-play conversations and coaching.

Here's the workflow: describe a scenario in plain English and you get a working voice or text simulation (16 languages) that embeds in Storyline, Rise, or any LMS. The AI evaluates the learner against rubrics you define, and then the scores, transcripts, and results report back to the course through SCORM and xAPI. There's a Storyline 360 variable bridge, so a sim can update Storyline variables like score, pass/fail, or even a character's mood to change states mid-course. And voice mode is what most people gravitate towards: let your learners talk through a hard conversation out loud so that the practice feel real.

Where it wins: when you want full control over your simulations and a practice layer inside the eLearning you already build, instead of moving your whole program to a new platform. Where it falls down: it launched in 2026, so it's the youngest tool on this list, and pricing is usage-based AI credits, which means budgeting per conversation rather than per seat. The Starter plan, built for portfolio-builders, is $9/month ($6/month paid annually), and it includes free hosting for your eLearning projects, which is why a lot of aspiring IDs use it for portfolio projects built around scenario-based learning.

Custom GPTs and Gemini Gems

The free DIY option, and where plenty of people should start. A well-prompted Custom GPT or Gemini Gem makes a surprisingly good role-play partner for low-stakes practice. Many devlin.ai users started exactly here before hitting the wall: the bot lives outside the course you're building, learners have to leave your training to use it, and nothing (no score, no transcript, no completion) flows back to the LMS. If you just need conversation practice and don't need it tracked or embedded, start here for free.

Yoodli and Second Nature

Two dedicated role-play platforms worth knowing about, with a caveat: I haven't run either through my 10-minute test, so I won't pretend to rank them. Yoodli focuses on speech and communication coaching as its own destination; Second Nature focuses on sales pitch practice for revenue teams. The fit question is platform versus layer: if you want a standalone coaching destination your learners visit, a platform-first tool may suit you; if you want practice inside the courses you already build, that's the layer approach. More on the format tradeoffs in AI role-play simulations.

Which AI tools help you build custom eLearning and automations?

Two options here, both from Anthropic: Claude Design if you want to start building this week without the technical friction, and Claude Code if you have a developer leaning and want the full power tool. Claude Code is the breakout tool in this category. It turns instructional designers into builders, and it has quietly replaced most of my other tools.

As of mid-2026, I've dropped ChatGPT and browser-based Claude for virtually all of my projects in favor of Claude Code: eLearning design and development, web design, automation development, document work with PDFs, and building devlin.ai itself. It's the best version of every other tool because it has persistent memory and every project gets its own dedicated workspace and dev container. For power users, cmux is the multiplier: it lets you run multiple Claude Code containers at once, and I'm often running 5 to 10 of them across different projects and codebases. Model-wise, I used all of the Opus 4 series models as they released, and Fable 5 is now my pick for long-running coding tasks.

When I started building devlin.ai, I assumed enterprise scale would require a security team, a compliance department, and multiple engineers just to survive procurement. I've been building it as a solo founder with Claude Code. Whatever ceiling you're assuming for a "designer who codes a little," it's higher than you think.

If "terminal-based" is where I lost you, start with Claude Design instead. Anthropic launched it in April 2026, and it has grown rapidly in popularity with instructional designers ever since, for a simple reason: it's a visual canvas rather than a command line, and what it produces is working HTML, CSS, and JavaScript. That's the same front-end output that tools like Storyline publish, which means the interactions and prototypes it builds are, functionally, eLearning. Describe what you want, watch it take shape, export a standalone HTML file, and SCORM-wrap it if it needs to live in your LMS. It's included with paid Claude plans (Pro and up), so if you're already paying for Claude, there's nothing new to buy. For a sense of what's possible, watch what Sabrina, an instructional designer on our team, built with Claude Design in a day: a software simulation with AI-powered customer calls and a portfolio website.

Where it falls down: it's front-end only and less capable than Claude Code, with none of the persistent workspaces or automation, so developer-leaning designers (myself included) will prefer Claude Code and its higher ceiling. But the two are designed to work together: when a design is ready to become a real build, Claude Design packages a handoff bundle you can pass straight to Claude Code. If you want to get up and running quickly and see what's possible with AI for designing learning experiences, Claude Design is the best place to start.

On automations: I built an AI operating system that lets AI draft content and help us make decisions across my businesses. And I think this pattern will shape the solutions instructional designers propose over the next few years. But you don't need to go that far to start. Zapier or n8n plus an AI step covers most L&D automation needs: course completion follow-ups, feedback summarization, content refresh reminders, for example.

One caution on the vibe-coding trend. SCORM-wrapping an AI-generated HTML interaction and dropping it into your LMS works, and practitioners are doing it. Just answer the maintenance question first: who fixes that code in 18 months when a browser update breaks it and you've changed jobs? Custom code you can't maintain is technical debt that will need to be paid eventually (unless your team is up to speed with a modern toolkit).

Which AI tools help with assessment and quality control?

Use AI to draft rubrics, generate question banks, and act as a second pair of eyes on your work. Never make it the final reviewer.

The drafting side is easy: paste your objectives and source content into Claude or ChatGPT and ask for scenario-based questions with plausible distractors, then cut the weak ones. Same for rubrics: AI produces a solid first draft in minutes that would take an hour by hand. The review side is where discipline matters. A useful prompt is to paste your storyboard alongside your objectives and ask the model to flag misalignment; it catches drift that's hard to see in your own work.

But keep a human at the end of the line. ATD's AI in Instructional Design report found that 20 percent of instructional designers avoid AI entirely, mostly citing lack of trust in AI-generated content, with copyright and accuracy concerns close behind. Those concerns are legitimate. The fix isn't avoiding AI; it's a hard rule that nothing AI-generated reaches a learner without human review. AI is your first-pass reviewer and your fastest drafter. You are the final reviewer, always.

What's the best AI tool stack for instructional designers?

A solo ID needs about 3 subscriptions. A team needs a different 4. An aspiring ID building a portfolio can get started for under $10 a month.

For reference, here's what I actually pay for as of July 2026: the Claude Max plan (which covers Claude and Claude Code), Magnific for media, and Articulate 360 with its AI Assistant. I still pay for ChatGPT, but it's mostly personal use now; Claude took over the work side.

Job	Solo / freelance ID	L&D team	Aspiring ID (portfolio)
Analysis, writing, storyboarding	Claude Pro	Your org's approved assistant (Copilot, Gemini, or Claude)	Claude or ChatGPT free tier
Authoring	Articulate 360 + AI Assistant	Articulate 360 + AI Assistant	Articulate 360 trial while you build
Media (images, voice)	Magnific, or skip it if 360's built-ins clear your bar	Magnific + ElevenLabs	Free tiers + 360 built-ins
Practice simulations	devlin.ai (Starter or Core)	devlin.ai (Team plan, usage-based credits)	devlin.ai Starter at $9/mo, with devlin.host portfolio hosting
Custom builds and automation	Claude Design to start, Claude Code when you outgrow it (both come with your Claude plan)	Claude Code + Zapier or n8n	Claude Design if you're on a paid Claude plan; otherwise learn prompting first
Rough monthly cost	~$30 to $60 plus your Articulate seat	Varies with seats and usage	From $9/mo plus free tiers

What's the best free AI tool for instructional design?

The free tier of a frontier assistant. My pick is Claude's free tier with ChatGPT's free tier a close second. A free frontier assistant covers analysis, learning objectives, storyboard drafts, assessment questions, and feedback on your writing, which is most of the thinking work in this field.

Know what "free" actually costs, though: usage caps that hit right when you're mid-project, no reliable project memory, and output that lives outside your course. The moment AI needs to work inside your deliverable (authoring, scored practice, tracked results), free general tools stop being the answer.

Frequently asked questions

Will AI replace instructional designers?

The evidence points to augmentation, not replacement, a question I cover in more depth in my honest answer on whether AI will replace instructional designers. In ATD's AI in Instructional Design report, 70 percent of instructional designers said AI improves the quality of their course design and 37 percent said it greatly reduced design time. Those are the numbers of a workforce getting faster and better, not disappearing. The parts of the job AI can't do (needs analysis with real stakeholders, judgment about what not to build, accountability for outcomes) are the parts that were always the hard parts.

Are general AI tools enough, or do I need specialized ones?

General tools cover the thinking work: analysis, objectives, storyboards, drafts, and review. Specialized tools win wherever the output has to live inside your course, which is why the authoring and simulation categories exist. A reasonable rule: buy a specialized tool only when your general assistant's output requires manual rebuilding to get into your deliverable.

Which AI tools do instructional designers actually use?

Per Dr. Philippa Hardman's survey research, 84 percent of instructional designers have used ChatGPT, and ATD's 2025 report found 80 percent of IDs use AI tools overall, almost entirely generative AI. My own paid stack as of July 2026 is Claude Max (including Claude Code), Magnific, and Articulate 360 with its AI Assistant, with devlin.ai for practice simulations.

What AI skills matter more than the tools?

Tools churn monthly, but skills compound. I've been making this argument, and the five AI skills I'd bet on haven't changed: writing effective prompts, mastering the tools you've picked, teaching other people about AI, reading and writing JavaScript (less relevant by the day with how good AI is at this), and building AI-powered automations. Notice that four of the five survive any individual tool dying, and the fifth (tool mastery) transfers because the tools keep converging on similar patterns.

If you're transitioning into the field and want structure while you build those skills, that's what Peck Academy is for. It's my licensed career school, and the training covers foundational ID theory, Storyline 360, and AI skills, including an AI-integrated flagship project built with extensive feedback.

I'll stand by this: this is one of the best times to be an instructional designer. The designers who treat AI as a way to accelerate and extend their capabilities, rather than a threat or a toy, are the ones proposing solutions nobody could think of three years ago. Pick your 3 or 4 tools, run everything new through the 10-minute test, and go deeper with how instructional designers are using AI in 2026.