How We Use AI to Plan Our Course Without Exposing Our IP

A Real Workflow for Creative Business Owners Who Can’t Afford to Expose What They’ve Built

Dis­clo­sure: Adobe has com­pen­sat­ed us to share our views on how we use AI tools, includ­ing Adobe Acro­bat Stu­dio, in our cre­ative process. All opin­ions are our own. We do not pro­mote tools we don’t believe in and where a tool has lim­i­ta­tions, we’ll tell you.

The AI Privacy Problem Nobody Talks About When Building a Course

If you’re build­ing a course and using AI to help plan it, the real ques­tion isn’t “which tool is the most pow­er­ful?”. It’s “Which tool can I trust with my intel­lec­tu­al prop­er­ty? And won’t be used to train AI.” This is a con­cern that many of us face.

Most gen­er­al-pur­pose AI tools—ChatGPT, Claude, Jasper—are pow­er­ful for:

• Gen­er­at­ing content
• Brain­storm­ing ideas
• Refin­ing messaging

But they are not designed to be the safest home for your pro­pri­etary frame­works or unpub­lished mate­r­i­al. Gen­er­al-pur­pose AI tools are opti­mized for output—not for inten­tion­al­ly pro­tect­ing your sen­si­tive con­tent. This guide doc­u­ments exact­ly how we nav­i­gat­ed that ten­sion while plan­ning our Hue Intel­li­gence course and how Adobe Acro­bat func­tioned as the secure core system—where pro­pri­etary con­tent is ana­lyzed, struc­tured, and protected.

Our Hue Intel­li­gence course is some­thing we have been devel­op­ing for a long time, build­ing on years of expe­ri­ence deliv­er­ing cours­es with LinkedIn Learn­ing for near­ly 167,000 learn­ers. Our course is a deep-dive into how to strate­gi­cal­ly use col­or and nav­i­gate design trends—not just for aes­thet­ics, but as a busi­ness tool that shapes per­cep­tion, dri­ves audi­ence behav­ior, and dif­fer­en­ti­ates brands. And because it rep­re­sents years of pro­pri­etary research, doc­u­ment­ed frame­works, and orig­i­nal method­ol­o­gy, pro­tect­ing it dur­ing the plan­ning process was crucial.

What we land­ed on was­n’t a sin­gle tool—it was a delib­er­ate deci­sion about which tool was trust­ed with which type of con­tent. Adobe Acro­bat Stu­dio was our core trust­ed sys­tem. Claude and Jasper on the oth­er hand, were exter­nal tools that han­dled select copy for refinement—nothing pro­pri­etary went into either plat­form. Here’s exact­ly how that work­flow func­tioned, what each tool actu­al­ly did, and what we’d do dif­fer­ent­ly next time.

The Core Rule That Changed Everything

Before we used a sin­gle tool, we defined one rule and gov­ern­ing prin­ci­ple: the sen­si­tiv­i­ty of the con­tent deter­mines the tool—not con­ve­nience or capa­bil­i­ty. This sounds sim­ple, but it required us to make explic­it deci­sions upfront about bound­aries that most peo­ple nev­er think to define until some­thing goes wrong.

How to Choose the Right AI Tool (Based on Content Sensitivity)

• Use Adobe Acro­bat Stu­dio → for inter­nal doc­u­ments, course con­tent, and pro­pri­etary IP
• Use Claude → for audi­ence analy­sis using san­i­tized notes
• Use Jasper → for mar­ket­ing and exter­nal content

Once we defined this rule, the result was a three-tool stack, each with a clear­ly defined role and a defined lim­it on what it could see.

Key Insight

The safest AI work­flows iso­late pro­pri­etary con­tent from gen­er­al-pur­pose AI tools.

Instead of upload­ing every­thing every­where, we:

• Kept core IP inside Acrobat
• Export­ed only what was safe
• Used oth­er tools downstream
  

What Adobe’s Data Policy Actually Says—And What It Doesn’t

This is the sec­tion that mat­ters most if you’re mak­ing a real deci­sion about where your course con­tent lives. We looked up the offi­cial pol­i­cy before writ­ing a sin­gle word of this guide, because over claim­ing on data pri­va­cy is a cred­i­bil­i­ty risk we’re not will­ing to take.

Here’s what Adobe offi­cial­ly states: Adobe does not train LLMs on your con­tent dur­ing your inter­ac­tions with Acro­bat AI Assis­tant. Prompts you pro­vide do not mod­i­fy the under­ly­ing mod­el. Data processed by AI Assis­tant is tem­porar­i­ly cached for up to 12 hours, then auto­mat­i­cal­ly delet­ed from Adobe cloud stor­age. Third-par­ty providers—including Microsoft Azure OpenAI—are con­trac­tu­al­ly pro­hib­it­ed from man­u­al­ly review­ing or train­ing their LLMs on Adobe cus­tomer data.

Here’s the hon­est caveat: No tool is 100% legal­ly bul­let­proof for all use cas­es. For high­ly reg­u­lat­ed indus­tries or enter­prise-lev­el con­fi­den­tial IP, your orga­ni­za­tion’s legal or com­pli­ance team should review any AI tool’s data pro­cess­ing terms before use. For most inde­pen­dent course cre­ators and small cre­ative stu­dios, Adobe’s stat­ed pol­i­cy offers the strongest data pro­tec­tion among tools in this work­flow. And it’s the rea­son Acro­bat Stu­dio earned our trust for inter­nal doc­u­ments when oth­er tools didn’t.

The Four Stages Where Acrobat Studio Did the Heavy Lifting

This is the sec­tion most blogs skip—the actu­al mechan­ics of what we did, in what order, and what hap­pened when things did­n’t work as expect­ed. We’re doc­u­ment­ing it in full because the hon­est ver­sion is more use­ful than a pol­ished summary.

Why Adobe Acrobat Is the Secure Core of This Workflow

In this work­flow, Adobe Acro­bat isn’t just anoth­er tool. It’s the con­trolled envi­ron­ment where sen­si­tive con­tent is ana­lyzed, struc­tured and pro­tect­ed.

The key dif­fer­ences are:

• It works only from the doc­u­ments you upload

• Does not train on your content

• Pro­vides page-lev­el cita­tions for every output

• Keeps source mate­r­i­al isolated

Unlike gen­er­al AI tools:

• No exter­nal data mixing

• No hal­lu­ci­nat­ed sources

• No expo­sure risk from prompts
  
  

Stage 1: Building the PDF Space and Uploading our Documents

To get start­ed, we went into Adobe Acro­bat Studio—accessible both on the web and via the desk­top app. We cre­at­ed a PDF Space and gave it a name that would be easy to track across our team. Before doing any­thing inside the plat­form, we rec­om­mend hav­ing every­thing you plan to upload already saved and orga­nized in one fold­er. The process is faster and less chaot­ic when you’re not hunt­ing for files while the inter­face is open.

We had 39 doc­u­ments to upload. That includ­ed PDF hand­outs we planned to use in the course, strat­e­gy doc­u­ments, years of col­or trend research we had doc­u­ment­ed our­selves, sev­er­al blogs we had writ­ten, copy from our email mar­ket­ing cam­paigns, info­graph­ics, and sta­tis­ti­cal data we had col­lect­ed over time. If we had read all of those doc­u­ments manually—even doc­u­ments we wrote ourselves—that would have been hun­dreds of pages, eas­i­ly sev­er­al hours of work just to re-famil­iar­ize our­selves with what we had.

You can upload up to 100 doc­u­ments, texts, PDFs, images, and links from Drop­box, Google Dri­ve, and more. One thing we noticed imme­di­ate­ly: Acro­bat had no trou­ble read­ing any of the file for­mats we uploaded. We have noticed that Claude and Jasper will some­times have issues read­ing PDF files—Acrobat nev­er did. That alone saved us time we would have oth­er­wise spent con­vert­ing or refor­mat­ting files.

Once all 39 doc­u­ments were uploaded, we could access key insights, lis­ten to an AI-gen­er­at­ed pod­cast from our files, or share the PDF Space with oth­ers for review. Acro­bat’s AI Assis­tant gen­er­at­ed ini­tial prompt sug­ges­tions based on what it had already start­ed analyzing—which gave us an imme­di­ate pre­view of what pat­terns it was pick­ing up across our content.

The Pod­cast Feature—This One Sur­prised Us
An under­rat­ed but gen­uine­ly impres­sive fea­ture is the AI-gen­er­at­ed pod­cast. After upload­ing our doc­u­ments, Acro­bat AI Assis­tant cre­at­ed two pod­cast files: a short ver­sion (around 8 min­utes) and an extend­ed ver­sion (almost 12 min­utes). These fea­ture two peo­ple hav­ing a real con­ver­sa­tion about the con­tent you uploaded—not a robot­ic sum­ma­ry, an actu­al dis­cus­sion about the mate­r­i­al. For peo­ple who don’t like to read through hun­dreds of pages, hear­ing your own work processed and dis­cussed in this for­mat was, hon­est­ly, mind-blowing.

We clicked through each pod­cast to access the tran­script, and any­thing that stood out—a pat­tern we had­n’t con­sid­ered, a con­nec­tion between doc­u­ments we had­n’t made—we copied and saved as a note direct­ly inside the platform.

Stage 2: Getting Started with a Custom AI Assistant

After the ini­tial upload, we start­ed click­ing through the sug­gest­ed prompts AI Assis­tant was sur­fac­ing to see what it was pick­ing up on across our 39 doc­u­ments. This was a use­ful ori­en­ta­tion step—it helped us iden­ti­fy pat­terns we had­n’t expect­ed and gave us a start­ing direc­tion before we knew exact­ly what ques­tions to ask.

Some of the sug­gest­ed prompts Acro­bat sur­faced on its own included:

• Extract cul­tur­al con­sid­er­a­tions linked to col­or psy­chol­o­gy in branding
• Sum­ma­rize emo­tion­al impacts of Peach Fuzz, Viva Magen­ta, and Veri Peri hues
• Devel­op a col­or strat­e­gy test frame­work based on pro­vid­ed best prac­tice exercises
• For­mu­late inves­tiga­tive ques­tions about cul­tur­al impacts on col­or meanings
• Pro­vide exam­ples of cul­tur­al col­or mean­ings and their brand implications
• Explore how col­or trends reflect cul­tur­al shifts and emo­tion­al needs

These were gen­uine­ly use­ful start­ing points—but they were still broad. Our goal was­n’t a gen­er­al sur­vey of col­or psy­chol­o­gy. We want­ed an AI part­ner that could look at every­thing we had built, syn­the­size our spe­cif­ic approach, and help us iden­ti­fy what top­ics to cov­er in the course. We need­ed some­thing more aligned with our voice and methodology.

Build­ing Our Own Cus­tom AI Assistant

Acro­bat gives you a choice of pre-built assis­tants: the default AI Assis­tant, Ana­lyst, Enter­tain­er, Instructor—or you can build your own. We test­ed the Ana­lyst agent first because the course con­tent is data-heavy, and it pro­duced bet­ter-aligned results than the default. But we quick­ly real­ized we want­ed some­thing even more spe­cif­ic to how we think and work.

So, we built our own cus­tom assis­tant. And here’s where Claude came in—not to see our course con­tent, but to help us write the con­fig­u­ra­tion prompt. We gave Claude a very spe­cif­ic con­ver­sa­tion­al request:

“Hey Claude, I want to cre­ate my own AI Assis­tant for Adobe Acro­bat Stu­dio. Based on what we do at NicteCreativeDesign.com, help me cre­ate the con­text I should pro­vide to the assis­tant so they can help me with a strat­e­gy that leans on our col­or exper­tise, brand design, and cul­tur­al communication.”

We also gave Claude a para­me­ter to keep the out­put under 1,000 characters—the lim­it for Acro­bat’s cus­tom assis­tant con­fig­u­ra­tion field. With­in min­utes, we had a cus­tom AI Assis­tant inside Acro­bat Stu­dio built around our spe­cif­ic brand context.

The dif­fer­ence was imme­di­ate. With our cus­tom assis­tant active, the sug­gest­ed ques­tions and the qual­i­ty of out­puts shift­ed to reflect our actu­al areas of exper­tise rather than gener­ic col­or the­o­ry top­ics. The assis­tant start­ed sur­fac­ing things like “for­mu­late inves­tiga­tive ques­tions about cul­tur­al impacts on col­or meanings”—and more impor­tant­ly, it was using our own con­tent and case stud­ies as the exam­ples in its answers, with­out us hav­ing to prompt it to do so.

One UI detail worth call­ing out: unlike Claude or Jasper, where uploaded ref­er­ence doc­u­ments get buried in a long chat thread, Acro­bat’s PDF Space keeps your files per­ma­nent­ly vis­i­ble in the left pan­el while you work. You can see all 39 doc­u­ments at any time, click on any spe­cif­ic file, and move between your chat and your source mate­r­i­al with­out los­ing con­text. For a research-heavy process like this, that’s not a minor convenience—it’s a fun­da­men­tal work­flow difference.

Stage 3: Annotating, Citing, and Building the Course Outline 

With our cus­tom assis­tant active and our 39 doc­u­ments ful­ly uploaded, we moved into the most inten­sive part of the process: using Acro­bat to ana­lyze our con­tent strate­gi­cal­ly and begin build­ing the course outline—without ever shar­ing the actu­al con­tent with a gen­er­al-pur­pose LLM.

Review­ing that vol­ume of mate­r­i­al is cum­ber­some, no mat­ter what tool you use. But what made Acro­bat’s approach dif­fer­ent was the abil­i­ty to stay con­nect­ed to the source while work­ing. Every time the AI Assis­tant pro­duced an out­put that felt like it should become a course note, a mod­ule top­ic, or a key insight, we could ver­i­fy exact­ly where it came from—and then save it.

Sav­ings Notes Direct­ly From the Chat

When we found respons­es in the chat that con­tained key infor­ma­tion we want­ed to pre­serve, we added them direct­ly as notes inside the PDF Space. Those notes then become acces­si­ble in the left-hand pan­el along­side all our uploaded doc­u­ments. What we appre­ci­at­ed was how Acro­bat han­dled the export: notes can be down­loaded not just as plain text, but as Word doc­u­ments, PDFs, or share­able links. This meant we could take our Acro­bat-gen­er­at­ed insights and, lat­er, upload those san­i­tized notes to Claude or Jasper for exter­nal copy work—without ever expos­ing the source doc­u­ments themselves.

Using Claude to Build a Bet­ter Out­line Prompt
As we got deep­er into the process, we want­ed to move from explor­ing our con­tent to actu­al­ly gen­er­at­ing a course out­line. At this point, we jumped back to Claude—not to share our doc­u­ments, but to help us craft a prompt that would get Acro­bat’s AI to pro­duce the most use­ful pos­si­ble output.

The prompt we land­ed on, which we then entered into our cus­tom AI Assis­tant inside Acrobat:

“Before cre­at­ing any­thing, ask me 2–3 clar­i­fy­ing ques­tions to help you build the best out­line pos­si­ble. Then, using only the doc­u­ments I’ve uploaded, cre­ate a pro­fes­sion­al course out­line for adult learn­ers that is self-paced. The course should teach how to use col­or strategically—not just for fleet­ing trends, but under­stand­ing what works, why it works, and how to apply it with inten­tion across brand and busi­ness con­texts. The con­tent should reflect expert-lev­el think­ing on col­or as a strate­gic tool, not a sur­face-lev­el design choice.”

The result was more of what we need­ed: a struc­tured course out­line gen­er­at­ed entire­ly from our own uploaded mate­r­i­al, ground­ed in our spe­cif­ic method­ol­o­gy, with no ran­dom sources or gener­ic filler content.

The Cita­tion Feature—Why This Changes Everything

The sin­gle biggest rea­son we used Acro­bat for this stage—and why it gen­uine­ly can’t be repli­cat­ed by Claude or Jasper in this context—is the cita­tion sys­tem. Every piece of con­tent Acro­bat AI Assis­tant ref­er­ences in its out­puts tells you exact­ly what doc­u­ment it came from and what page. Not a URL from the inter­net. Not a syn­the­sized claim with no ver­i­fi­able ori­gin. The spe­cif­ic file and the spe­cif­ic page inside your PDF Space.

For a course that ref­er­ences years of research, doc­u­ment­ed case stud­ies, and orig­i­nal data, this mat­ters enor­mous­ly. It meant we could trace every mod­ule top­ic, every strate­gic claim, and every frame­work sug­ges­tion back to our own source material—and ver­i­fy it was actu­al­ly ground­ed in what we had built, not what the AI had invented.

Anno­ta­tions and Team Collaboration

One fea­ture that Claude and Jasper fun­da­men­tal­ly can­not offer: the abil­i­ty to anno­tate the actu­al doc­u­ments you’ve uploaded. Inside Acro­bat PDF Spaces, you can open any uploaded file and add anno­ta­tions, com­ments, and notes direct­ly to the doc­u­ment itself. This was use­ful for flag­ging sec­tions of our own hand­outs that need­ed updat­ing, mark­ing pages we want­ed to ref­er­ence in spe­cif­ic course mod­ules, and not­ing where graph­ics need­ed to be created.

It also opens up a team col­lab­o­ra­tion lay­er. We were able to invite team mem­bers to com­ment on spe­cif­ic anno­ta­tions or add notes about graph­ic require­ments direct­ly with­in the workspace—all tied to the source doc­u­ments, not float­ing in a sep­a­rate project man­age­ment tool.

One les­son we’d pass on: when you switch between assis­tant types in Acrobat—for exam­ple, mov­ing from your cus­tom assis­tant to the built-in Ana­lyst to com­pare outputs—some of the sug­gest­ed prompt ques­tions that Acro­bat sur­faces may dis­ap­pear and not reap­pear when you switch back. We lost a few use­ful sug­ges­tions this way. Our rec­om­men­da­tion is to take screen­shots of any sug­gest­ed prompts you want to revis­it before switch­ing assis­tants, or ask the AI to gen­er­ate a saved list of rec­om­mend­ed tasks you can access at any time.

Stage 4: Deepening the Analysis with Out Custom Assistant 

By this stage in the process, our cus­tom assis­tant had become the most pro­duc­tive part of the entire work­flow. The dif­fer­ence between using a pre-built assis­tant and one con­fig­ured around our spe­cif­ic brand con­text, col­or exper­tise, and com­mu­ni­ca­tion approach was significant—not just in the rel­e­vance of out­puts, but in the qual­i­ty of the ques­tions it prompt­ed us to consider.

Where the built-in Ana­lyst agent pro­duced use­ful but broad results, our cus­tom assis­tant start­ed sur­fac­ing insights that were direct­ly tied to our own doc­u­ment­ed approach. When it sug­gest­ed we “for­mu­late inves­tiga­tive ques­tions about cul­tur­al impacts on col­or mean­ings,” it was focused on pulling insight from our con­tent that we had uploaded. And it used that mate­r­i­al as an exam­ple in its answer, not just as a citation.

This is the dis­tinc­tion that sep­a­rates Acro­bat Stu­dio from every oth­er tool we test­ed for this stage: with Claude, Jasper, and even Per­plex­i­ty, you have to explic­it­ly and repeat­ed­ly prompt the AI not to pull from the inter­net and to ref­er­ence only your uploaded documents—and even then, it isn’t always guar­an­teed. With Acro­bat PDF Spaces, the sys­tem is built to work only from what you’ve shared. We nev­er had to remind it.

By the end of Stage 4, we had a set of notes cov­er­ing poten­tial mod­ule top­ics, course struc­ture options, key ques­tions the course need­ed to answer, and con­tent gaps we had­n’t pre­vi­ous­ly iden­ti­fied. All of it was ground­ed in our own doc­u­ment­ed exper­tise, cit­ed back to spe­cif­ic pages in spe­cif­ic doc­u­ments, and saved in for­mats we could move into the next stage of the workflow.

Where Claude and Jasper Fit In

We want to be clear about some­thing: using Claude and Jasper at the copy and mar­ket­ing stage was­n’t a fallback—it was a delib­er­ate choice made because both tools are gen­uine­ly great at what they do. But we delib­er­ate­ly drew a bound­ary when it came to access­ing our IP.

Claude: Audience Analysis and Gap Identification

Claude is use­ful to help you with:

• Audi­ence analysis

• Iden­ti­fy­ing gaps

• Refin­ing messaging

Once we had our ini­tial course out­line built inside Acro­bat, we knew we need­ed to refine it fur­ther. But since we did not trust shar­ing our orig­i­nal course con­tent or the out­line itself with Claude, we export­ed our saved notes from Acro­bat, which includ­ed the syn­the­sized insights, not the source doc­u­ments, and uploaded the result­ing Word doc­u­ment to Claude instead.

We asked Claude to do two things with that mate­r­i­al: first, cre­ate a detailed audi­ence per­sona we could use to tar­get our mar­ket­ing con­tent; and sec­ond, ana­lyze the out­line for any gaps that might cre­ate con­fu­sion or unan­swered ques­tions for course tak­ers. Claude is built on ana­lyt­i­cal process­es but is also more conversational—you can bounce ideas and test pos­si­ble direc­tions in a way that feels more like work­ing through a prob­lem with a col­lab­o­ra­tor than query­ing a data­base. That qual­i­ty made it the right tool for this stage.

What Claude nev­er saw: the course out­lines them­selves, the hand­out drafts, the research PDFs, the strat­e­gy doc­u­ments, or any of the pro­pri­etary frame­works doc­u­ment­ed in our Acro­bat Space. The bound­ary held throughout.

Jasper: Sales Copy, Social Posts, and Team Tracking

Jasper is effec­tive for generating:

• Sales copy

• Social posts

• Brand voice consistency

Once Claude had pro­duced the audi­ence per­sona and gap analy­sis, we uploaded the output—along with the refined out­line notes—to Jasper. In this stage, we didn’t trust shar­ing our pro­pri­etary source mate­r­i­al with Jasper, and only shared our curat­ed con­tent results from Adobe Acro­bat Stu­dio. From there, we asked Jasper to cre­ate sales page copy for our web­site and a set of social media posts to tease and pro­mote the course.

The rea­son we use Jasper at this stage specif­i­cal­ly is: Brand Voice. We’ve trained Jasper on our spe­cif­ic tone, lan­guage, and com­mu­ni­ca­tion style, so the copy it pro­duces sounds like us—not like gener­ic AI out­put or one that we con­tin­u­al­ly have to train. For exter­nal-fac­ing con­tent that needs to match our brand pre­cise­ly, that train­ing pays off.

Jasper also func­tions as a track­ing and col­lab­o­ra­tion tool for our team. When the copy is ready for review or needs fur­ther edit­ing, we man­age that sta­tus inside Jasper’s Can­vas, which keeps every­one aligned with­out requir­ing a sep­a­rate project man­age­ment sys­tem for content.

Pricing Overview: What This Workflow Actually Costs

You can start this work­flow for free and upgrade only where your vol­ume demands it. Here’s the cur­rent pric­ing for every tool in this stack.

What Makes This Workflow Different

Most cre­ators will:

• Upload every­thing into one AI tool

This work­flow:

• Sep­a­rates think­ing from protection
• Iso­lates sen­si­tive data
• Con­trols what each tool sees

Frequently Asked Questions

These are the ques­tions course cre­ators and cre­ative busi­ness own­ers ask most often when they start think­ing about AI tools and IP protection.

Is it safe to upload course con­tent to Chat­G­PT or Claude?

It depends on your plan set­tings and your risk tol­er­ance. Both Ope­nAI and Anthrop­ic offer options to opt out of using your con­ver­sa­tions to train their models—but the defaults vary depend­ing on your plan type, and it’s worth read­ing the spe­cif­ic terms for your account. For unpub­lished IP, pro­pri­etary frame­works, or con­tent that rep­re­sents a sig­nif­i­cant busi­ness asset, a doc­u­ment-spe­cif­ic tool like Adobe Acro­bat Studio—where data is processed and delet­ed with­in 12 hours and not used for mod­el training—is a mean­ing­ful­ly more con­ser­v­a­tive choice. This isn’t about Claude or Chat­G­PT being bad tools. It’s about match­ing the tool to the sen­si­tiv­i­ty of what you’re sharing.

Does Adobe Acro­bat use my uploaded doc­u­ments to train its AI?

Accord­ing to Adobe’s offi­cial pol­i­cy, no. Adobe states that AI Assis­tant does not train LLMs on cus­tomer con­tent dur­ing inter­ac­tions, that prompts do not mod­i­fy the under­ly­ing mod­el, and that processed data is tem­porar­i­ly cached for up to 12 hours before being auto­mat­i­cal­ly delet­ed from Adobe cloud stor­age. Third-par­ty providers used in the infrastructure—including Microsoft Azure OpenAI—are con­trac­tu­al­ly pro­hib­it­ed from man­u­al­ly review­ing or train­ing their mod­els on Adobe cus­tomer data. For most inde­pen­dent course cre­ators and small stu­dios, this rep­re­sents the strongest avail­able data pro­tec­tion in a stan­dard AI work­flow stack. For high­ly reg­u­lat­ed indus­tries, we’d rec­om­mend review­ing the full pol­i­cy with your legal team.

What is PDF Spaces and how is it dif­fer­ent from just using AI Assis­tant on a sin­gle document?

PDF Spaces is a col­lab­o­ra­tive work­space inside Acro­bat Stu­dio that lets you com­bine up to 100 files, links, images, and doc­u­ments in a sin­gle shared envi­ron­ment. When you use AI Assis­tant inside a PDF Space, it ana­lyzes all of those doc­u­ments together—not just one at a time. For course plan­ning specif­i­cal­ly, this means you can have research papers, draft out­lines, hand­out drafts, and your own strat­e­gy doc­u­ments all in one space, with the AI draw­ing con­nec­tions across all of them simul­ta­ne­ous­ly. The left-pan­el inter­face also keeps all your files vis­i­ble while you work, so you’re nev­er search­ing for a doc­u­ment mid-session.

Can I use the free ver­sion of Acro­bat AI Assis­tant to plan a course?

The free tier gives you 5 life­time requests—enough to get a sense of the fea­ture, but not enough for a full course plan­ning process. The $9.99/month add-on (avail­able after June 30, 2025) is the prac­ti­cal entry point for ongo­ing use, giv­ing you 1,000 AI requests per month. Acro­bat Stu­dio at $24.99/month adds PDF Spaces and Adobe Express Premium—if you’re build­ing a course and plan­ning to cre­ate accom­pa­ny­ing con­tent assets, the Stu­dio tier is a bet­ter val­ue for the full workflow.

Do I need a sep­a­rate Acro­bat sub­scrip­tion if I already have Cre­ative Cloud?

Not necessarily—but it depends which CC plan you have. Most Cre­ative Cloud plans include basic Acro­bat func­tion­al­i­ty, but AI Assis­tant and PDF Spaces are not includ­ed in all plans. Before pur­chas­ing Acro­bat Stu­dio sep­a­rate­ly, log into your Adobe account and check what’s cur­rent­ly bun­dled. The pric­ing page at adobe.com/acrobat/pricing lets you com­pare plans with your cur­rent sub­scrip­tion in view.

What’s the dif­fer­ence between Acro­bat Pro and Acro­bat Studio?

Acro­bat Pro ($19.99/month) cov­ers core PDF edit­ing, cre­ation, form fill­ing, and e‑signatures—everything you’d need for stan­dard doc­u­ment work. Acro­bat Stu­dio ($24.99/month) adds AI Assis­tant, PDF Spaces, and Adobe Express Pre­mi­um to that foun­da­tion. For course cre­ators who want to use the AI-pow­ered plan­ning work­flow described in this guide, Stu­dio is the rel­e­vant tier. If you only need doc­u­ment edit­ing and sign­ing, Pro is sufficient.

I already use Notion or Google Docs to plan my course. Why would I use Acro­bat instead?

You don’t have to choose—these tools serve fun­da­men­tal­ly dif­fer­ent func­tions. Notion and Google Docs are excel­lent for gen­er­al project man­age­ment, task track­ing, and col­lab­o­ra­tive writ­ing. Acro­bat Stu­dio’s advan­tage is spe­cif­ic: when you have PDF documents—research papers, pub­lished reports, your own hand­outs, data exports—that you want to ana­lyze with AI with­out upload­ing them to a gen­er­al-pur­pose LLM, Acro­bat is the tool built for that job. Think of it as a com­ple­ment to your exist­ing plan­ning stack, not a replace­ment. We use both.

Can I share my PDF Space with my team?

Yes—PDF Spaces sup­ports team col­lab­o­ra­tion. You can invite oth­ers to view, com­ment on, and anno­tate doc­u­ments with­in the space. Team mem­bers can add notes direct­ly to uploaded files, flag sec­tions that need updates, and coor­di­nate on con­tent with­out need­ing to down­load and re-upload doc­u­ments. One impor­tant caveat: if you share your PDF Space via the QR code gen­er­at­ed for the pod­cast fea­ture, the entire space is shared—not just the pod­cast. Be inten­tion­al about what’s in the space before shar­ing it externally.

Key Takeaways

1. AI work­flows should be designed around trust boundaries—not just effi­cien­cy. Adobe han­dles the pro­tect­ed con­tent. Oth­er tools han­dle curat­ed out­puts. This dis­tinc­tion is what pro­tects your IP.
   
   
2. Adobe Acro­bat Stu­dio’s AI Assis­tant ana­lyzes only the doc­u­ments you share with it, ref­er­ences only your inter­nal sources in its out­puts, and deletes processed data with­in 12 hours. For inde­pen­dent course cre­ators with pro­pri­etary con­tent, this makes it the most pro­tec­tive tool in a stan­dard AI work­flow stack.
   
   
3. Claude and Jasper remain gen­uine­ly valuable—at the right stage. The key is estab­lish­ing a clear bound­ary before you start about what each tool is and isn’t allowed to see, and main­tain­ing that bound­ary con­sis­tent­ly through­out the process.
   
   
4. The PDF Spaces fea­ture is one of the most under­uti­lized capa­bil­i­ties in Acro­bat. Com­bin­ing up to 100 doc­u­ments in a sin­gle AI-enabled work­space, with per­sis­tent file vis­i­bil­i­ty and team anno­ta­tion tools, fun­da­men­tal­ly changes how you can ana­lyze and build on your own exist­ing content.
   
   
5. You can start test­ing this work­flow for very lit­tle. Five free Acro­bat AI requests, Claude’s free tier, and Jasper’s 7‑day tri­al cov­er your first com­plete course plan­ning ses­sion at min­i­mal cost. Upgrade the spe­cif­ic tool where you hit lim­its first.
   
   
6. Build­ing a course is an IP event. Treat­ing your plan­ning tools with the same inten­tion­al­i­ty you’d apply to a client contract—knowing exact­ly what each plat­form does with your data—is no longer option­al for seri­ous con­tent creators.
   
   

What We’re Building Next

The work­flow doc­u­ment­ed in this guide is the same one we used to build our upcom­ing Hue Intel­li­gence course—a deep dive into how to strate­gi­cal­ly use col­or and nav­i­gate design trends, not just for aes­thet­ics, but as a busi­ness tool that shapes per­cep­tion, dri­ves behav­ior, and dif­fer­en­ti­ates brands in a crowd­ed market.

We’re launch­ing the course very soon. If you want to be the first to know when it’s avail­able and get ear­ly access pric­ing, grab our free col­or trends guide to get started.

Final Thought

AI didn’t replace our process. It made it more structured.

The real advan­tage came from know­ing where not to use cer­tain tools—not just where to use them.