When we released VideoGPT a few months ago, it enabled viewers to ask questions while watching the video and receive answers based on the video’s transcript and on-screen information.
Shortly after launch, many of you shared feedback asking for more control, specifically, a way for VideoGPT to respond to viewers’ questions that are “not directly discussed” in the video itself.
We listened to that feedback and, as a result, introduced VideoGPT Custom Prompts.
VideoGPT Custom Prompts bridge this gap by allowing you to predefine answers to questions your viewers or students are likely to ask—especially those that cannot be answered directly from the video content.
These may include general FAQs, site or content inventory references, related video recommendations, course team details, or product and service information.
Use Cases for VideoGPT Custom Prompts
VideoGPT custom prompts help organizations handle questions that go beyond the content shown in a video. They enable consistent answers to common viewers’ questions, guide users to relevant material, suggest related videos for deeper learning, share information about the people or context behind the content, and optionally surface additional courses or services.
These use cases ensure viewers receive clear guidance without interrupting the instructional flow or relying only on the video content.
1. General FAQ
Viewers regularly ask general questions that sit outside the main video flow—such as clarifications of certain terms or keywords, prerequisite knowledge, or broad concepts assumed by the course. Repeating these explanations across videos would interrupt pacing and reduce focus.
Custom prompts let you define clear answers to these common questions once, and have VideoGPT use them consistently wherever the question appears. This ensures viewers get reliable explanations when they need them, while keeping the video content focused and uninterrupted.
For example, A student may ask general questions while watching a lesson. In such cases, you can define a custom prompt for VideoGPT to control how these questions are answered.
The custom prompt could be:
|
“When viewers ask for basic definitions, background concepts, or general study guidance that is assumed knowledge, respond using the predefined explanations provided below.” Query 1: If a student asks, What is a variable in programming? VideoGPT responds with a clear explanation: “A variable is a named container used to store data values that can change while a program is running. Variables are important because they allow programs to work with dynamic data instead of fixed values. For a detailed introduction, refer to the Introduction to Programming section, where variables are explained step by step.” Query 2: If a student asks, How should I practice this concept? VideoGPT responds with predefined guidance: “Practice this concept by working through 5–10 examples on your own. Start with the guided exercises in the workbook, then try creating your own variations. Most students spend about 2–3 hours practicing before feeling comfortable with this topic. For more structured practice, refer to the ‘Practice Exercises’ section of the course.” |
2. Site Inventory
Organizations often host a large collection of related videos, pages, or resources within a single site. Viewers may explore this content in different ways—skipping around, revisiting topics, or accessing content out of sequence. As a result, they often ask questions that reference related materials or background information available elsewhere on the same site inventory.
Custom prompts allow VideoGPT to recognize these questions and respond by pointing viewers to the relevant modules, lessons, or sections already available in the site inventory, without attempting to re-teach the content inside the current video.
The custom prompt that you can write could be:
|
“When viewers ask how one topic relates to another, respond to them by following inventory items and timestamps:” Core Programming Concepts → Recursion Basics (timestamp 03:15) Data Structures → Trees and Graph Traversal (timestamp 06:30) Do not explain or summarize the content. Refer only to the lesson titles and timestamps listed above.” Query: If a viewer asks, “How does recursion relate to data structures?” while watching a video on algorithm design. VideoGPT responds: “Recursion is introduced in the Recursion Basics lesson under Core Programming Concepts at 03:15. Its application within data structures is covered in the Trees and Graph Traversal lesson in the Data Structures module, starting at 06:30.” |
3. Suggesting Related Videos
In some cases, viewers may want to explore additional content beyond the video, such as related features or supporting videos.
Custom prompts allow you to define when and how VideoGPT should suggest these related items. The suggestions are based only on the information you explicitly provide in the custom prompt, ensuring accuracy and control.
For example, a viewer may want to continue learning beyond the video by exploring related or supporting content. In such cases, you can define a custom prompt for VideoGPT to guide how recommendations are made.
The custom prompt that you can define could be:
|
“When viewers ask what to watch or learn next, respond using only the following predefined resources: Module: Advanced Programming Related videos: • Object-Oriented Programming (OOP) Present these as optional next steps. Do not explain the content in detail and do not suggest resources outside this list.” Query: If a student asks, “Where can I learn about Data Structures?” VideoGPT responds by recommending predefined resources, such as: “You can explore this topic further in the Advanced Programming module. As optional next steps, you may want to watch the videos on Object-Oriented Programming (OOP), Data Structures & Algorithms, and Design Patterns, which are listed in the next section.” |
4. Course Team Information
Students may want to know more about the people behind a course to better understand the expertise and perspective shaping the content. These questions can relate not only to the instructor, but also to other contributors such as faculty members, course designers, reviewers, testers, or subject-matter experts involved in creating or validating the material.
This information is usually not included in lesson videos, as videos focus on teaching specific topics. Custom prompts allow VideoGPT to respond with accurate, predefined details about the people involved in the course.
For example, a student may want to know more about the people involved in creating or delivering a course. In such cases, you can define a custom prompt for VideoGPT to control how these questions are answered.
The custom prompt that you can define could be:
|
“When viewers ask about the people involved in this course, respond only using the following predefined course team information: Lead Instructor: Dr. Alex Morgan — Lead instructor with 10+ years of experience teaching computer science, specializing in software architecture and data structures. Bio: https://example.com/alex-morgan Subject-Matter Expert: Jane Patel — Senior Software Engineer who contributed to course design and technical accuracy. LinkedIn: https://linkedin.com/in/janepatel Academic Review Team: Internal academic reviewers who validated the course content for accuracy and clarity. Do not add, infer, or speculate beyond this information” Query: If a student asks, “Who is the instructor talking now?” or “Who helped design or review this course?” VideoGPT responds using the predefined information, such as: This section is presented by Dr. Alex Morgan, the lead instructor for this course. Dr. Morgan has over 10 years of experience teaching computer science and specializes in software architecture and data structures. You can view Dr. Morgan’s full bio here: https://example.com/alex-morgan The course content was developed with input from Jane Patel, a Senior Software Engineer who contributed to the course design and technical accuracy. You can view her LinkedIn profile here: https://linkedin.com/in/janepatel |
5. Product or Service Promotion
Viewers may ask questions related to additional learning resources, advanced courses, or premium features that require a subscription or paid access, or even pricing. These questions may arise while they are exploring content and looking for ways to continue or deepen their learning.
Custom prompts allow you to predefine how VideoGPT should respond to such questions. These responses can introduce relevant courses, subscriptions, or premium learning options, along with information about available discounts, promotional offers, or access codes, in a subtle and supportive manner, without disrupting the learning experience.
For example, a student may ask about continuing their learning or accessing more advanced content while watching a lesson. In such cases, you can define a custom prompt for VideoGPT to guide how these questions should be answered.
The custom prompt that you can define could be:
| “When viewers ask about additional or advanced learning options, respond only using the following predefined information:
Course name: Advanced Programming Course Includes: 12 modules, hands-on exercises, and assessments Access: Available via paid subscription Course page: https://example.com/advanced-programming Promo code: VIDEOGPTCOUPON (30% discount) Clearly explain what the viewer gains by upgrading. Mention the promo code only if relevant. Do not add pricing details or suggest content not listed here.” Query: If a student asks, “Is there a full course available for this topic?” VideoGPT responds with predefined details, such as: “Yes. This topic is part of the Advanced Programming Course, which includes 12 modules, hands-on exercises, and assessments designed to deepen your understanding. Access to the full course is available through a paid subscription, which also unlocks additional learning resources and practice materials.” You can view the full course details here: https://example.com/advanced-programming If you decide to enroll, you can use the promo code VIDEOGPTCOUPON at checkout to receive a 30% discount.” Using a coupon is a great way to track how engaging and converting VideoGPT is, as you can track coupon usage in your eCommerce system. |
6. Sales & SDR-Style Engagement
In some scenarios, organizations want VideoGPT to actively support sales conversations in addition to providing informational answers.
In other words, you can instruct VideoGPT to take on a light Sales Development Representative (SDR) role. In this mode, VideoGPT responds to viewer questions by highlighting relevant product benefits, asking qualifying questions, and guiding viewers toward the next step, such as exploring pricing, requesting a demo, or starting a trial.
For example, a viewer may ask a general product-related question while watching a video. In such cases, you can define a custom prompt for VideoGPT to guide how these questions should be handled.
The custom prompt could be:
| “You are acting as a light SDR for the company. When users ask general product questions, explain how the product addresses common use cases related to their query. Ask one simple qualifying question to better understand their needs, and suggest a relevant next step such as viewing pricing, starting a trial, or contacting sales. Keep the response informative and non-aggressive.”
Query: If a viewer asks, “What can this product help me with?” VideoGPT responds by explaining how the product supports common use cases related to the question, asking a simple qualifying follow-up such as “Are you looking to use this for individual or team use?”, and then gently suggesting an appropriate next step, such as reviewing pricing details, starting a trial, or contacting the sales team for more information. |
Enabling VideoGPT on Your Videos
Before using VideoGPT custom prompts, you must ensure that your videos have transcripts available and are published to VideoGPT. VideoGPT works by using video transcripts, which can be generated automatically or uploaded manually.
You can enable this in two ways:
- For individual videos: You can generate or upload transcripts and publish them to VideoGPT directly from the video editor. This is useful when you want to enable VideoGPT for specific videos only. For step-by-step instructions, refer to this guide.
- For all videos (global settings): You can configure auto-transcription and publishing options at the account level using the global Auto-Transcribe settings. This ensures that transcripts are automatically generated and published to VideoGPT for all videos you upload going forward. For more details, refer to this guide.
Steps to Configure VideoGPT Custom Prompts
You can configure VideoGPT custom prompts from your Cincopa account settings.
Step 1: Log in to your Cincopa account, navigate to the profile section, and click the drop-down menu.
Step 2: From the drop-down, select Account Settings.
Step 3: In the Account Settings page, click Global Settings from the left-hand sidebar.
Step 4: In Global Settings, search for VideoGPT Custom Prompts.
Step 5: Enter the information you want VideoGPT to display when viewers ask questions that are out of scope for the video.
Step 6: Click Save.
