Jenni.ai is an AI academic writing tool built on top of GPT. Within two years, it has grown from 0 to 3 million users and $6M in ARR, with a team of ~ 8 people (as of August 2024).
How did they achieve this? What can we learn from Jenni.ai’s success?
I’ll break down the lessons into two parts: Build + Sale.
PART 1. BUILD
Lesson 1. Target a real problem.
Jenni.ai’s target users are students who have a paper deadline approaching. Imagine that they need to submit their 2,000-word paper by 11:59 PM, tonight. And it’s 9PM, they haven’t started a thing and they are panicking.
Jenni.ai has many TikTok videos of the target user (TikTok Link)
Because this is a real, urgent problem, users will be crying tears of joy for a legit solution:
Lesson 2. GPT is suitable to solve this problem.
The First Rule of Student Papers: they need to *look like* a paper. Do student papers need some original thoughts or groundbreaking ideas? No.
GPT is great at imitating patterns — churning out something that “looks like” a paper.
For instance, I entered the following title as a prompt: “when llm agents will replace humans”.
This is what I got from Jenni.ai. It looks legit. And the legitness is so visceral that I felt nauseated as a recovering academic.
A few things stand out:
- It creates a two-part title. And the first part, “The future of work”, makes me feel the weight of the paper.
- The prose is appropriately dry, and the sentences are somewhat penetrable upon close examination. Nailed the academic writing style.
- It has citations. The four citations show the author’s diligent research and thorough understanding of the subject.
Therefore, I’d say it *looks like* a paper 100%.
The exception is for more serious academic writing (e.g. grants, peer-reviewed journal papers) where the content will be under close scrutiny by Reviewer 2. The bullshit might be called out. Citations (or lack of certain citations) will be complained about.
For instance I checked the reference, and some of them are not relevant to the content, which would be noted in a rigorous peer review.
Despite the blemishes, Jenni.ai will help students with an approaching deadline to churn out 2000 words that look like a paper.
Lesson 3. What unique value does Jenni.ai offer compared to ChatGPT?
A challenge for all GPT-wrapper products: If GPT can solve the problem, why would users want this product instead of ChatGPT?
Jenni.ai solves this in two ways:
(1) Friction-free workflow
Just like GitHub Copilot for coding, Jenni.ai autocompletes the next sentence in writing.
For a writer under time pressure, Jenni.ai’s seamless workflow beats the back-and-forth of ChatGPT.
(2) Tailored to the specific use case (academic writing)
A significant issue with ChatGPT is its tendency to hallucinate — it often generates references that don’t exist.
Jenni.ai addresses this by using RAG (search academic papers first then feed into GPT) to provide real, verifiable citations. These citations can be exported in different styles (e.g., APA 7).
This feature alone will drive users to choose Jenni.ai over ChatGPT.
(3) Additionally, it can throw some AI magic dust to the draft: paraphrase, make longer, or write opposing arguments.
Again, my inner voice:
So far, Jenni.ai has created a compelling AI product to solve a real problem. The next question is how to sell?
Part 2: Sale
Lesson 4. GTM strategy: reach out to users where they are.
David Park, Jenni.ai’s CEO, shared their go-to-market strategy in this tweet: “How to go from 0 to $5M ARR profitably (step by step)” (link)
TLDR: Reach out to users where they are. For Jenni.AI, it’s creating viral TikTok videos.
I highly recommend reading this tweet, and in fact, all of David Park’s tweets. His content is crisp, fresh, and informative.
Lesson 5. Monetization: use a 3-Step formula
Jenni.ai employs a classic 3-step formula to encourage users to pay:
Value + Suspension + Discount
Here’s what a typical user journey looks like:
Step 1: Value: Quick time-to-value
User inner voice: “As a freemium user, it only took me 5 minutes from signup to my first AI-generated content. It clicks. I get it. I am excited to write more. I am on my way to get my paper done!”
Step 2: Suspension
After the initial experience, users will face the following suspensions with 5 mins:
Type 1: Free word limit
User inner voice: “However, I’ve hit my free word limit for the day. Come on, I have a paper due today, I can’t wait until tomorrow!!”
Type 2: Blurred references
User Inner voice: “What, this reference section is blocked? It’s almost there. I can almost see it!”
A successful series of suspensions will lead to the following (and that’s exactly what I felt):
User Inner Voice: “The suspension is creating this itch in me that makes me want to pay!”
Step 3: Discount
The discount is the homerun that brings users to the checkout.
User inner voice: “But I don’t want to pay… There are two voices arguing in my mind. One insists: ‘The price is too high!’”
Suddenly: “A wild 10% off coupon appears. You know what? I’ll just pay it! Yay!”
This strategy is highly effective. It’s extracting payment from users, yet leaving them satisfied with their decision.
(Yes, I paid! Purely in appreciation of their excellent onboarding + monetization flow)
Lesson 6. Understand Users’ Deepest Fears
For a sophisticated AI product like Jenni.ai, its mission statement is uncharacteristically … vague.
To me, it shows exactly the reason behind Jenni.ai’s success:
It truly understands its users and their fears.
While using ChatGPT to paraphrase is ok, using it to write entire academic essays is cheating. Such academic misconduct can lead to serious consequences (e.g. failed class or probation).
As a result, users face two concerns:
(1) They don’t want to view themselves as cheaters.
(2) They don’t want to be caught cheating.
Jenni.ai’s vague mission statement allows students to use the tool without explicitly confronting concerns about academic integrity.
Jenni.ai also mitigates the “cheating” concern through omissions. For instance, it avoids messages like “guaranteed undetectable!” Such statements would only heighten users’ awareness of academic integrity issues.
It even goes so far as to tell its influencers in affiliate program:
“Don’t Portray Jenni.ai as a cheating tool (you can position Jenni as “secret weapon” or “study hack” though!)”
Is it good or bad practice? Curious what you think!
If you are an AI product builder, I’d recommend you to experience Jenni.ai. It provides textbook examples of how to onboard users, how to monetize, and how to create a delightful user workflow.
New challenge: How to deal with copycats
Imitation is common in tech. As Ex-Google CEO Schmidt said:
Say to your LLM, ‘Make me a copy of TikTok. Steal all the users, steal all the music, put my preferences in it, produce this program in the next 30 seconds, release it, and in one hour, if it’s not viral, do something different along the same lines.
For Jenni.ai, its functions and workflow are hard earned, but can be easily stolen. A copycat product can copy Jenni.ai’s UI + workflow, while offering it at a much cheaper price.
I am curious to see how Jenni.ai will address this challenge as the product becomes more visible, more profitable, and more coveted. This is a common problem for all GPT-Wrapper products.
Thanks for reading till the very end. Follow me for more content like this.
Is there any AI product you’d like to see in my case study? Please comment below!
#AIProductCaseStudy
