AI Creator Tools · Kajabi
Creator Studio
Breaking the marketing barrier for creators by turning content they already have into full marketing funnels. Kajabi’s first monetized AI-generation product.
Christine Noh · Lead Product Designer · Feb 2023 – Apr 2024 · Pre-beta · Beta · GA
Why it mattered
Earn = stay
Creators who never earn were Kajabi’s largest churn risk. Expertise didn’t become
contacts or revenue.
- According to our Marketing Experience Survey (Feb 2022, n=578),
~60% weren’t highly confident in marketing their business - It took creators 8+ hrs to build a funnel
- At 3.8% monthly churn, most of it came from that unprofitable segment
the unprofitable segment
Creator Studio was a bet that marketing help on content they already had would move activation and retention.
- Prior Kajabi AI already moved activation, but 3–10 word prompts
capped personalization. - Creator Studio started from content creators already had — roughly
12,000 words of expertise per hour of course video — with 10.5k on
the waitlist before Beta launch.
My role
Lead Product Designer · 0→1 through GA · team of 11
1 designer · 1 PM · 1 PMM · 2 researchers · 5 eng · 1 QA
Owned
- End-to-end UX
- Assistive AI principle
- First AI Design System patterns
Co-led
- Strategy + funnel pivot
- Research direction
- First AI NRR pricing
Constraints
- No AI Design System precedent
- Pictory-capped stack (editing & unit price)
- Mid-flight re-scope
First AI Design System patterns
- AI intake — one flow for video, audio, blog, PDF, web page, and YouTube — without prompt-engineering
- Generation states — loading, progress, success, error, and retry for long-running AI jobs
- Async toasts — “3 of 4 files ready” with a CTA back to the project
- Editable AI outputs — every highlight/output can be added, removed, or edited
- Media-type picker — select from 37+ generated output types
Later reused across Kajabi’s AI surfaces.
01 Pre-beta
01 Pre-beta · Research findings
Marketing was the barrier,
not content creation
- Marketing was the hard part (Marketing Experience Survey · Feb 2022 · n=578)
- Creators wanted help from content they already had (AI Content Creation Survey · Nov 2022 · n=356)
- They were already paying a stack outside Kajabi, ~9 tools on average, spending roughly $1,300/mo on extras (Third Party Tool Usage · Aug 2023 · n=534)
01 Pre-beta · Strategy
Repurpose existing video
for social media content
Pre-beta let creators select their existing long-form videos from courses or coaching sessions. Then use Creator Studio’s AI to auto-select recommended parts of the video — with the ability to add or remove what AI had selected — and export the selected portion into a new clip fitted for various social media platforms.
- Opening bet — prove “creators will repurpose content they already have”
- White-label Pictory — partner on an MVP by white-labeling instead of building the video pipeline, so we could validate in weeks
- ChatGPT API for highlight selection — AI recommends sections from the video and cuts them to fit the selected platforms
- Platform set — IG, FB, TikTok, LinkedIn, YouTube — most used platforms by Kajabi creators
- Trade-off — less control over the engine, faster thesis test
01 Pre-beta · design
Repurpose existing video
for social media content
User experience: pick an existing course or coaching video → AI drafts a social cut → creator edits → resize for the platform → publish to Facebook or Instagram.
Design decisions
- Reuse existing lessons — access already created content instead of starting from scratch
- Editable AI highlights — AI recommends a draft so starting isn’t a blank-page barrier; creators still control what stays in the clip
- Light resize editor — edit to fit the platforms where creators connect with their audience
- In-product syndication — export and re-upload is a drop-off; posting in-flow closes the loop
01 Pre-beta · User feedback
Give me the whole funnel
I ran creator sessions after pre-beta. Repurposing was validated, but a social clip was too small a slice of the problem. What surfaced next reframed the product:
- Give me the whole funnel, not just a clip — matched earlier demand: 61% wanted AI support designing funnels, automations, and pricing.
- Let me pull from YouTube — best footage already lived there; that finding drove YouTube import in Beta.
- Make my course footage easy to find — fast search inside Creator Studio.
- Give me enough editing to drop outside tools — richer in-product editing so creators can cancel third-party subscriptions and work entirely inside Kajabi.
- Move the business, not just the workflow — profit and activation KPIs, not only clip-editing time.
02 Beta
02 Beta · Strategy
Roadmap
- Team direction — helped reframe the empty funnel as the business problem so we aligned on full-funnel assets
- Build the full funnel — turn one source video into a marketing system that helps creators earn and stay
- Pricing bet — AI usage scales cost, so we tested willingness to pay in invite-only beta — before GA — and opened a new revenue line
- Trade-off — more scope and complexity; aimed at the business problem, not only the editing one
-
Pricing tiers
- Free — 1 video/mo included, so pricing didn’t kill activation
- Basic · $9/mo — unlimited content from 4 videos/mo
- Premium · $29/mo — 30 videos/mo + Adobe Express Premium
02 Beta · design
Expand from one clip
to the entire funnel
User experience: select a video (YouTube link, upload, or course/coaching session — with search) → choose media types → AI generates assets → optionally clip video → publish across the funnel.
Design decisions
- 37+ media types — build the entire funnel from one video — emails, landing pages, social, and more
- YouTube import — paste a link to use footage that already lives where creators publish
- Course search — find the right lesson in seconds instead of scrolling the library
- Pricing model — unlock more AI usage with an optional add-on instead of upgrading the whole Kajabi plan
02 Beta · User feedback
More editing control
Beta expanded what creators could make, but the editing, feedback, and cost model needed work.
“Within 2–3 hours, I created a year’s worth of blog posts for my video blog. It was slick, it was professional.”
Beta creator
- More editing control — richer video editing than the light editor
- Tell me when it’s ready — clear status when 37+ media types finish generating
- The vendor cost too much — Pictory was too expensive to scale
03 GA
03 GA
Rebuild the stack underneath: new vendors, a real editor
Beta asked for more editing control, clear generation status, and a cost model we could scale — so GA rebuilt those three layers.
- Switch to Picsart — lower unit cost at scale and widen generation capability
- Adobe Express editing — professional-grade control inside Creator Studio, beyond Beta’s light editor; the integration also opened opportunities and integrations across other parts of Kajabi
- Generation status messaging — toasts (and optional email) keep creators informed as assets finish
After GA
Outcome
Descript, Canva, CastMagic
Beta survey · Dec 2023 (n=68)
- 57% found Creator Studio easy or very easy
- 51% said their business would be somewhat or very negatively affected without it
- 43% fully or partly replaced 1+ external tools
Measured: paid adoption, assets generated, outside-tool replacement in beta. Aimed at $650k annualized NRR.
Not claimed: churn, contacts, or GMV from Creator Studio alone.
Future: prove generation → contact → purchase; judge revenue by usage, not only add-on purchase.
What's next
Once creators can build the funnel, how do we make it work automatically?
- Auto-syndication & scheduling (Activation + NRR) — publish/schedule via Adobe SDK
- Performance analytics (Retention + GMV) — which assets convert
- Multi-language repurposing (NRR) — one asset → funnels in multiple languages
- Remove filler words (NRR) — strip ums, uhs, dead air
- Bulk create projects (NRR) — e.g. 10 lessons → 10 projects
- Video to audio (NRR) — podcast-ready audio from the same source
Reflection
What worked · what I’d change
What worked
- Trusting rich inputs creators already had
- Listening after pre-beta (YouTube import and the clip → funnel pivot)
- Treating pricing and vendor economics as design problems
- Buying and partnering (Pictory first,
Adobe Express at GA) let us validate fast and still ship professional editing.
What I’d do differently
- Validate the funnel bet earlier — would have saved a release cycle
- Test editing-control needs earlier — we replaced the light editor with Adobe Express at GA
- Create dashboards to see insights from week 1: activation, completion, repeat use
Appendix
My thoughts on AI
- A tool, not a replacement. AI is a great tool — and it should stay a tool, not a replacement for the human brain. Anything the model wrote or highlighted had to stay editable, because AI was never meant to be the final content producer. The person stays the author and makes the final call.
- Best for getting started. AI is most helpful when it helps you do something you don’t know how to do yet — getting information, exploring ideas, and getting a first draft started.
- Strong at language tasks. It’s genuinely useful for writing text, summarizing context, capturing the main idea, and analyzing data.
- Weaker at complex, detail-heavy work. It’s less helpful on complex tasks that require many details.
- Words are the friction. AI has to be instructed with words, but sometimes an action is much faster than putting it into words — and clarifying what to do, and what not to do, can take a lot of extra words. Good AI products shouldn’t force people to become prompt writers: they should use the context people already have, offer guided choices, and let people edit directly.
- It can be confidently wrong. AI can sound right and still be wrong, so a human has to verify what’s true.
- Only as good as its input. A generic prompt in gets generic output out; a creator’s own long-form content in gets output in their voice.
- AI drafts, humans judge. AI can offer drafts, options, and speed — but it doesn’t know your audience, your business, or your taste. The human still brings the judgment, the voice, and the responsibility for what ships.