Why course creators need an AI chatbot
Course creators sell on a strange page. The sales page for a $499 course has to do three jobs at once: convince a stranger the course is for them, answer every objection they could possibly have, and close the sale on the spot. Most sales pages do the first two with long scrolling copy. The third happens, or does not happen, in the visitor's head.
The gap between "convinced this is interesting" and "clicked Buy" is the most expensive gap in the creator economy. Reading the sales page costs the visitor 90 seconds. Reading every FAQ entry to find their specific concern costs another 90 seconds. By the time they have not found the answer to "does this work if I am a complete beginner" or "what is the refund policy" or "is week three really live or recorded," they have closed the tab. They were not unconvinced. They just ran out of patience before they ran out of objections.
A chatbot trained on your sales page, your curriculum, your FAQ, and your refund policy answers those objections in one sentence with a citation. It does this at 2am when you are asleep. It does this for the visitor scrolling on their phone who would never have read the FAQ accordion to the bottom. It does this in the language the visitor used in their head, not the language you used on the page.
This is what a chatbot does for course creators, and it is not the same job as a chatbot does for a SaaS docs site or an ecommerce store. The buying decision is more emotional, the price point is higher, the asynchronous "let me think about it" pattern is stronger, and the cost of a bounced evaluator is one missed $499 sale, not one missed $19 SKU.
The persona: "Lena Hoffmann" on Kajabi
Lena is a productivity coach. She has been online for six years, has a podcast with 80,000 downloads a month, an email list of 12,000 people on Kit (the platform that used to be ConvertKit), and a flagship course called Deep Focus for Knowledge Workers that she sells on Kajabi. Deep Focus is priced at $499 as a self-paced evergreen course and $899 as a cohort version she runs twice a year with weekly live calls.
She sells roughly 30 to 50 self-paced seats a month and 60 to 90 cohort seats during each cohort launch. Her annual revenue from Deep Focus is in the low six figures. She has one virtual assistant, Yara, who handles inbox triage for four hours a day. Everything else, including the sales page copy, the email sequences, the podcast, and the cohort calls themselves, is Lena.
Yara's inbox has a recognizable shape. About 40 percent of her time goes to existing-student support: password resets, video that will not play, how to get the workbook. Another 40 percent goes to pre-purchase questions from people who have read the sales page and want to ask one specific thing before they buy. The remaining 20 percent is everything else, podcast guests, sponsorship requests, and the occasional crisis.
The pre-purchase questions are the ones that bother Lena, because she knows that for every one Yara answers there are probably five visitors who had the same question, did not feel comfortable emailing a stranger, and just closed the tab. Those five are her leak.
Pre-purchase Q&A: the killer use case
The single most valuable thing a chatbot does on a course sales page is answer "is this course for me." This question takes 30 different forms. Is it for beginners. Is it for people who are not knowledge workers. Is it for parents. Is it for people with ADHD. Does it work if I am self-employed. Does it work if I am in a 9-to-5. Is it too basic if I have already read four books on focus. Is it too hard if I have not read any.
A long sales page cannot answer all 30 forms. A chatbot trained on the sales page, the curriculum, and Lena's "who is this for" essay can. Lena watches the chatbot transcripts for two weeks and sees a pattern: 38 percent of all pre-purchase questions are some form of "is this for me." Another 22 percent are "what does week three actually cover." Another 18 percent are refund-policy. Another 12 percent are pricing-shape, mostly "is there a payment plan." The remaining 10 percent are everything else.
The math on those 38 percent is the most important math in the creator economy. If Lena's sales page gets 4,000 monthly visitors, and 38 percent of pre-purchase questions are "is this for me," and even one in ten of those visitors had the question and bounced because the page did not answer it, that is 152 lost visitors per month. At her conversion rate of 1.5 percent, that is roughly 2 lost sales per month. At $499 a seat, that is $998 a month, or $12,000 a year, leaking from a single bucket of preventable bounces.
A chatbot on the sales page closes that bucket. Not all of it, because some of those visitors are tire-kickers who would not have bought anyway. But conservatively half. Lena's expected lift from pre-purchase deflection alone is in the range of one extra sale per week from the evergreen and two extra cohort seats per launch.
Module-content Q&A and the Teachable/Kajabi platform limits
Lena would also love a chatbot inside the course, helping students who are confused about a specific module. This is a real wish and a real limit.
Teachable and Kajabi both gate the course player behind login. Their custom-code areas (the Site Footer / Power Editor on Teachable, the Header Page Scripts on Kajabi) inject scripts on public pages and, in some cases, on the post-login student dashboard, but the course player itself is more restricted. Per Thinkific's documentation, Site Footer Code "will not execute on the default Performance Checkout" and behavior in the course player is similarly constrained. Kajabi's custom-code field injects into the head of every page on the website "except Offer checkout pages and upsell pages."
The honest answer for course creators is: ChatRaj runs reliably on your public sales page, your free-content blog, and your homepage. On the gated course player, it depends on the platform and the plan. For the persona this page is for, that is fine, because the public sales page is where the money is decided. Inside-course help is a nice-to-have; pre-purchase deflection is the unit economic.
Lead capture for waitlist and email list
Not every pre-purchase visitor is ready to buy tonight. For the cohort version of Deep Focus, the dynamic is even more pronounced: cohorts only run twice a year, so a visitor who lands between launches needs to be captured on the waitlist or they are lost.
ChatRaj's lead-capture flow handles this naturally. When a visitor asks a cohort-related question and the next cohort is months away, the bot answers the question, then offers: "The next cohort opens September 15. Want me to add you to the waitlist so you get the early-bird discount?" If they say yes, the bot collects the email, the lead fires through a webhook into Kit (or ConvertKit, Mailchimp, Beehiiv, ActiveCampaign), and the visitor enters Lena's nurture sequence.
For the evergreen course, the same flow captures lower-intent visitors who want to think about it. The bot offers a free "Focus Audit" PDF in exchange for email, the lead drops into a 5-email nurture sequence that ends with a soft pitch, and 4 to 8 percent of captured emails convert to sales within 30 days. This is the back half of the chatbot's job, and for most creators it is worth more than the immediate-conversion half.
Refund-policy Q&A: the most-asked question
If you sell a course, you already know that "what is your refund policy" is the most-frequently-asked pre-purchase question. It is the last objection. The visitor is mostly sold, but they want to know they can get out if it does not work.
The bad answer is to bury the refund policy on page 4 of a Terms of Service link in the footer. The medium answer is to put it in the FAQ accordion at the bottom of the sales page. The good answer is a chatbot that, when the visitor types "refund," responds in two sentences: "We offer a 14-day no-questions-asked refund. Email lena@example.com from the address you bought with and we process it within 48 hours."
That sentence, delivered at the moment the question is in the visitor's head, closes more sales than any amount of sales-page copy. Lena's transcripts show the pattern: 18 percent of pre-purchase questions are refund-related, and about two-thirds of those visitors purchase within the next 10 minutes of session time.
Cohort versus evergreen pricing questions
Deep Focus has two SKUs at two price points. The sales page tries to explain the difference. Real visitors still ask: which one is for me, can I upgrade from evergreen to cohort later, is the cohort discount available if I am already in evergreen, what is the payment plan on the cohort.
A chatbot trained on the sales page handles these well, because the answers are already on the page; the visitor just did not scroll to find them. The bot is faster than the page. For pricing-shape questions specifically, ChatRaj's confidence on these is high (the answers are short, factual, and one-paragraph), which means the bot rarely abstains and rarely hallucinates.
What ChatRaj does NOT do for course creators
Honest scope is the most important part of this page. ChatRaj is not a course platform. There are things creators sometimes assume a chatbot can do that ChatRaj does not.
ChatRaj does not handle membership gating. If a question requires the bot to know which course a logged-in student is enrolled in, ChatRaj cannot do that. The bot is anonymous and content-grounded, not authenticated and roster-aware.
ChatRaj does not handle Stripe or billing. The bot cannot tell a student what they were charged, when their next payment is due, or whether their card failed. Those questions go to your platform's billing dashboard, not to a chatbot.
ChatRaj does not integrate with LMS progress. The bot does not know which module a student finished, what their quiz score was, or whether they completed week three. The bot cannot say "you should review module two before moving on" because it cannot see module-two completion.
These limits are real and important to name. For Lena's actual job-to-be-done, which is closing the evening browser who is reading the sales page and almost-but-not-quite ready to click Buy, none of these matter. Pre-purchase deflection happens before login, before Stripe, before any LMS progress exists.
Setup on Teachable (custom code areas)
Teachable supports custom code through two surfaces, with caveats. The Code Snippets area in the school admin lets you inject scripts site-wide, and the Power Editor / Custom HTML blocks let you add HTML inline on specific pages. Per Teachable's own help center, "Custom HTML blocks are not enabled for new schools" by default; if your school does not have them, log in and contact Teachable Support to enable.
For ChatRaj, you only need site-wide injection, so the Code Snippets area is the right surface. Paste the script tag, save, and the widget appears on every public page including your sales pages. The bot does not appear inside the course player on the default plan, which matches what we want.
Setup on Kajabi (custom code areas)
Kajabi is the easier of the two. The Header Page Scripts field under Settings > Site Details accepts any HTML and JavaScript, and per Kajabi's docs, "Custom code added to the Header Page Scripts field will be included in the head tag of every page on your website, except Offer checkout pages and upsell pages."
That exception is fine: you do not want the chatbot on the checkout page (the visitor is mid-purchase, do not distract them). The Pro plan ships custom code out of the box. The Basic and Growth plans need Kajabi Access, which is a $99 per month add-on that unlocks the custom code editor. If you are on Basic and not ready to add Kajabi Access, the easier path is a single Custom Code Block on the sales page itself, which is available on all plans.
Setup on Thinkific (Pages with custom code)
Thinkific's Site Footer Code field, available on the Basic plan and higher, accepts the standard script tag. Per Thinkific's documentation, the footer code runs on most pages but "will not execute on the default Performance Checkout" and behaves differently in the Course Player. For sales pages, the footer code surface is the right one. The free plan does not include footer code; you need at least Basic.
For Podia, the Custom Code section in the site settings handles the same job. Paste, save, deploy.
ROI: two extra sales per week from "is this for me" deflection
Lena's first 60 days with ChatRaj look like this. The bot handles roughly 600 pre-purchase conversations per month. Of those, around 230 are some form of "is this for me," 130 are curriculum-specific, 110 are refund-related, 70 are pricing-shape, and 60 are everything else. The bot captures roughly 80 emails per month from visitors who were not ready to buy but wanted the waitlist or the lead magnet.
The immediate sales lift is harder to attribute cleanly, because some of those visitors would have bought anyway. The conservative attribution model, which counts only sales that followed a chatbot conversation by less than one hour, shows roughly 8 to 12 extra sales per month from the evergreen and 14 to 22 extra cohort seats per cohort launch. At Lena's price points, that is in the range of $5,000 to $8,000 of attributable monthly revenue from a tool that costs $29 a month and took 20 minutes to install.
The 80 captured emails per month feed a nurture sequence that converts at 4 to 8 percent, adding another 3 to 6 sales per month with a 30-to-60-day lag. The total program lift, after 90 days, is roughly two extra sales per week on a steady-state basis, with cohort launches adding bigger one-time spikes.
This is what pre-purchase deflection looks like for course creators. It is not the same job a chatbot does for SaaS docs or an ecommerce store. It is more emotional, more asynchronous, more sensitive to one specific question being answered at the moment the visitor has it. For creators on Teachable, Kajabi, Thinkific, or Podia, that job is what ChatRaj is built to do.