Thinker - AI Teaching Assistant FAQ

Thinker is your always-on course companion, designed to answer questions about your content, support your customers as they learn, and save you from replying to the same questions — helping you grow revenue without needing to expand your team.

Yes. When starting out, Thinker is only visible to Site Owners and Site Admins on your site. You control when it’s turned on for your learners. 

Currently, Thinker is enabled on a per-course basis and cannot be restricted to a specific group of learners within a course. If a learner is enrolled in a course where Thinker is active, they will see it. It will not appear in any courses you have excluded.

Yes. You can customize several aspects:

• Name and profile image: Choose how Thinker is displayed to learners.
• Greeting message: Customize the initial message learners see, including sample questions to guide them.
• Voice and tone: Personalize Thinker’s tone of voice to be playful, formal, or anything in between.

Yes. A learner’s conversation history is stored in a secure database and protected using the same security and privacy standards we apply across Thinkific. Both learners and customers can revisit past conversations at any time.

Yes, it has multilingual capabilities. For example, if the course content is in English, a learner can ask a question in Spanish, and Thinker will use the course content to provide an answer in Spanish.

The hallucination rate is designed to be very low. Thinker is heavily weighted to provide answers that are 100% factual and directly based on the creator's provided content. It is intentionally restricted from sourcing information from the wider internet to ensure accuracy and relevance.

Thinker currently supports PDF, text and on-platform video lessons.

Yes, you can upload individual PDFs that are not tied to your courses that Thinker can reference in its responses without being visible to learners. Learn more here: Setting Up Thinker.

Additionally, you can create a draft text, video or PDF lesson for this content. Saving this lesson as a draft allows Thinker to consume the content without it being visible to learners.  

Yes, Thinker is compatible with the new course player - if your site was created on or after January 28, 2026, you will have access to this feature. Learn more: Introducing Thinkific's New Course Builder 

Thinker, AI Teaching Assistant, is available for customers on all Thinkific Plus plans, with the number of available monthly credits depending on your plan. If you're already a Plus customer, reach out to your dedicated Customer Success Manager to learn more. Curious, but not yet on Plus? Discover how Plus can supercharge your business growth.

Thinker, AI Teaching Assistant, is available for customers on all Thinkific Plus plans, with the number of available monthly credits depending on your plan.

Each time Thinker answers a learner’s question, one credit is used. For example, if a learner asks Thinker, “Where can I find the workbook?” and Thinker replies with the answer, that uses one credit.

If you have reached your monthly credit limit, reach out to your Customer Success Manager to explore your options.

You can track their credit usage from the Thinker Analytics dashboard.

No - any unused credits within that monthly limit do not roll over.

Thinker can only be trained on content from your courses. Specifically, it can reference video (transcripts), text, and PDF lessons.

Not automatically. After adding new content to a course that's already accessed by Thinker, you must perform a manual sync from the teaching assistant Content access tab. This will re-index the course and incorporate the new material.

• It can use content from hidden or private courses, but only if you explicitly add that course to Thinker's knowledge base.
• It cannot use content from fully unpublished courses.
• It can use draft content within a published course. This allows you to add supplementary material (like an e-book) to a draft lesson for Thinker to use without making that lesson visible to learners.

Upload any general or support-type content that will help Thinker answer common learner questions, such as:

• Help center or FAQ articles
• Internal or external support documentation
• Policy and process documents
• General reference guides that don’t fit inside a single course

Thinker appears as a button in the course player for any course that it has access to. Clicking the button opens the chat interface alongside the course content, similar to the Discussions feature. Learners can also access it on the Thinkific or your Branded mobile app.

Yes! Learners can access Thinker on the Thinkific or your Branded mobile app versions 2.7 and above.

Thinker offers different content access methods, which determine how it references your synced content:

• Use all synced content: references any synced content, regardless of whether a learner doesn’t have access. 
  • Note: this option won’t give full access to the course, but will provide the learner with a URL that takes them to the course landing page if they don’t have access.
• Match user access: only references content the user has access to.

No. Thinker is disabled when a learner is taking a quiz or exam.

Thinker isn’t limited to the course a learner is currently viewing. It includes built-in context settings, so you can focus answers on a specific lesson or course when needed. At the same time, it can reference any content you’ve granted it access to.

No. Uploaded PDFs are private by default. Thinker uses them behind the scenes to improve its answers, but it does not show or link to the original files in the chat.

No. Uploaded PDFs are treated as global reference content. Thinker can use them to answer questions for any learner in any course where Thinker is enabled, regardless of their enrollment in a particular course.

Yes. A dedicated analytics page provides insights into:

• Total users and number of responses.
• An "estimated hours saved" metric.
• The most active users.
• A chart of message volume over time.

Yes. The analytics section includes a detailed message history, allowing you to review all conversations or filter them by a specific learner to see their interactions.

Yes. Every response from Thinker includes a thumbs up/thumbs down option. This feedback is tracked in the analytics, showing you the total number of positive and negative responses to help gauge Thinker's effectiveness.

You can track their credit usage from the Thinker Analytics dashboard.

Thinker does not access or store personally identifiable information (PII). It only uses the content from your Thinkific school (such as course materials) to answer learner questions inside your school.

Thinker stores learner chat history only for that learner, so they can return to previous answers. This history isn’t shared across schools or used to train any public or global model.

No. Thinker is not training a new model or creating a custom LLM. Thinker uses your synced course content to provide context-aware responses, but it does not change, tune, or train the underlying model. Thinker behaves the same for all schools and does not store or carry forward your proprietary content.

Thinker does not access or store PII. Learners do not need to provide additional personal details to use it; it does not require or handle sensitive, financial, health-related, or identity-specific information. If a learner includes that information in a conversation with Thinker, it is treated as part of their chat history and is not used for training or shared outside the learner’s school.

No. Thinker only uses your school’s content to answer learner questions within your school. It does not share your content with other customers, and it is not used to train or improve any global or shared AI models.

This depends on your organization’s internal privacy policies. Many large organizations conduct a DPIA for any new software — especially tools that interact with end users. Thinker itself does not process PII or create new models from your data, but you may still need a DPIA to meet your governance or compliance requirements.

Yes. School admins control whether Thinker is enabled or disabled for their school. If Thinker is turned off, no learner data is processed by it.

Disclaimer: This information does not constitute legal advice. 

We strongly recommend you consult with a legal professional to determine if you need to update your Terms and Conditions. You may need to inform your learners that you are using an AI tool to assist in answering questions and that their interactions are stored.

We provide several ways to gauge the quality of Thinker’s responses:

1. Thumbs up/Thumbs down buttons: Every response includes simple thumbs-up/down buttons for instant feedback.
2. Conversation analysis: Our systems analyze conversations to identify trends, measure overall satisfaction, and flag potential issues or feedback that learners mention in the chat. This helps us (and you) understand Thinker’s performance.
3. Feedback button: We have added a 'Provide Feedback' button, allowing learners to give more detailed, open-ended comments on their experience, with an opportunity to do so every seven days.

Thinkific has invested in extended jailbreak testing and has a number of safety procedures in place to ensure that Thinker stays on track and isn’t tricked into revealing information that it shouldn’t. This is an ongoing process that we will also continue to evolve along with the industry.

No. Thinker does not have direct access to your raw content files (e.g., video files, PDFs). It interacts with a secure index of your content's information, not the files themselves. It is designed to answer questions based on this index, not to provide bulk downloads or transcripts of your material.

Thinkific only works with companies that contractually agree not to train their models on customer content (such as AWS Bedrock, Google AI and Open AI). If any data is used to improve functionality at the platform level, it is anonymized or aggregated.

You retain full ownership of both the content you input and the content generated by Thinkific’s AI tools. We will never use your content (original or AI-generated) without your consent.

No, your course material will not be used to train third-party Large Language Models (LLMs). We take protecting your content very seriously and enforce this through binding legal agreements.

We only partner with leading AI sub-processors who are contractually bound to a zero-data-retention and zero-training policy. This means they are prohibited from using your data for training their models or for any purpose other than immediately generating a response to your student's query.

All interactions with AI tools on the Thinkific platform fall within the bounds of these strict data protection agreements.

Thinkific implements multiple protections to reduce the risk of harmful or inappropriate AI-generated content. This includes moderation filters and alignment with the OWASP Top 10 for large language model applications.

Access to your content is strictly controlled and limited. Here is a breakdown of who can and cannot access it:

• Your Content Storage: In order to prepare your content for use with Thinker, we pre-process, index and store it in a vector database. We built and manage this technology to ensure your data remains under our control, not on a third-party AI platform.
• AI Sub-processor Access: Our carefully vetted AI sub-processors (like Google, AWS, OpenAI) have temporary, programmatic access to the data required to generate an answer. This access is encrypted and lasts only for the duration of the request. They are contractually forbidden from storing or using this data afterward.
• Other Customers: Your content is never visible or accessible to any other Thinkific customer.
• Thinkific Employee Access: Access to your data is governed by a strict least-privilege policy. Only specific, authorized Thinkific engineering personnel can access the system for maintenance and troubleshooting, and they will never use your content for any other purpose.

Any data shared with sub-processors lives off-platform long enough to fulfill the request and satisfy short-term abuse monitoring, after which it is deleted.

Before a new AI partner is added, Thinkific performs security, privacy and legal reviews, signs Standard Contractual Clauses, and continuously monitors performance. If a new sub-processor is selected for use, Thinkific will provide all customers with 15 days' notice via email. 

Any data shared with sub‑processors lives off-platform long enough to fulfill the request and satisfy short‑term abuse monitoring, after which it is deleted.

See here for more information on Thinkfic’s sub-processors.

Yes - customers from regulated regions (EU, UK, California, etc.) have the right to object and, if necessary, cancel affected services within 30 days for a refund. Additionally, Thinkific provides at least 15 days' advance notice before adding new sub-processors.

Thinkific is committed to complying with GDPR in Europe, CPRA in California, GRPA in South Korea, Québec’s LAW‑25, Canada’s PIPEDA, and other applicable laws. We adhere to best practices for responsible AI and data use, protecting both you and your customers. 

For more information, see Data Processing Addendum.

Thinkific holds a SOC 2 Type 2 report (Security trust qualification) and undergoes quarterly third‑party penetration tests, and operates 24/7 SIEM monitoring. Courses are encrypted at rest and in transit, and accounts can use multi-factor authentication with least-privilege access. 

For more information, see Security Overview. 

Thinkific uses the following measures to protect data:

• Once your course is published, it will be encrypted at rest with AES-256 and TLS 1.2+ in transit.
• Your Thinkific account can be set up with multi-factor authentication and least-privilege access for all employees.
• You'll be set up with separate production and staging networks and will undergo quarterly penetration testing.
• Course administrators can be set up with multi-factor authentication for secure access to your course sites.

We use a multi-layered approach to ensure that your data remains secure when interacting with our AI partners. Our process is built on three key pillars:

• Rigorous Vetting: Before any partnership is established, all potential AI sub-processors must pass a comprehensive security, privacy, and legal review. This includes an assessment of their data protection practices, infrastructure security, and key certifications (such as SOC 2 and ISO 27001).
• Strict Contractual Obligations: We require all AI partners to sign a stringent Data Processing Agreement (DPA) that includes Standard Contractual Clauses (SCCs). These legally binding contracts explicitly prohibit them from using your data for their own purposes, including training their models, and enforce strict data handling and confidentiality requirements.
• Continuous Monitoring: Our commitment to security doesn't end after the contract is signed. We continuously monitor our partners' performance and compliance to ensure our high standards for security and privacy are met on an ongoing basis.

Yes - use of our AI tools is optional. If you opt not to use them, the AI will not access any of your data or content.