Reported impact
Business outcome
Khan Academy accelerated its roadmap for tutor-like experiences, gave students more personalized help, and opened a path for teachers to tailor instruction more quickly.
- Customer Experience
- Productivity Improvement
OpenAIEducationAI and Machine LearningGenerative AI / LLMCustomer ExperienceProductivity Improvement2023 OpenAIOA#10a37fEducationglobalKhan Academy needed a better way to personalize instruction because students learn at different speeds and teachers need scalable help to tailor support, explanations, and classroom materials for each learner.began testing the newest version of OpenAI's language model in 2022Mar 14, 20235GPT-4 · Khanmigo0 2023-03-14 5
Khan Academy
Khan Academy needed a better way to personalize instruction because students learn at different speeds and teachers need scalable help to tailor support, explanations, and classroom materials for each learner.
began testing the newest version of OpenAI's language model in 2022
thousands of lessons
Solution
Khan Academy uses GPT-4 to power Khanmigo, an assistant that acts as both a virtual tutor for students and a classroom helper for teachers. The nonprofit had long wanted tutoring flows that could ask individualized follow-up questions, push students to explain why they answered a certain way, and connect lessons to each learner's interests, but those experiences were difficult to build with traditional software. With GPT-4, freeform student prompts and lesson context enter Khanmigo, which returns conversational guidance, deeper conceptual questioning, and contextual explanations rather than just one fixed answer. On the teacher side, the same model layer is being tested for tasks such as drafting classroom prompts, creating materials, and summarizing how a class is doing on a given day. Khan Academy is piloting the system carefully, using early users to flag errors while it studies how responsibly to expand AI into learning workflows.
Data flow
Student questions and lesson context flow into Khanmigo, GPT-4 generates individualized tutoring exchanges and conceptual follow-ups, and the same model stack can also summarize classroom progress or draft teaching materials for educators.
Solution architecture
- Compute
- GPT-4 Provides the reasoning and conversational capability behind tutoring and classroom-assistant interactions.
- Serving
- Khanmigo Delivers student tutoring and teacher-assistance experiences on top of Khan Academy's education platform.
Architecture clues
- Khanmigo is being run as a limited pilot with error reporting from early users
- The same GPT-4 foundation is intended for both student and teacher workflows
- The workflow depends on freeform prompt understanding and conversational back-and-forth
Evidence from the source
Khan Academy is announcing that it will use GPT-4 to power Khanmigo
One of GPT-4's chief capabilities is being able to understand freeform questions and prompts
The nonprofit is testing out ways teachers could use GPT-4, such as writing classroom prompts or creating instructional materials
We're making sure students aren't just understanding how to do the problem, but really understanding the concepts behind them
Transparency Evidence & methodology
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
How scores are calculated
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Customer reference value71/100
- Problem clarity20/20Measures whether the customer's operational problem is specific enough to compare.
- Implementation depth25/25Rewards named components, data flow, integrations, and architecture detail.
- Outcome specificity16/20Measures whether the reported benefit is concrete rather than a generic claim.
- Quantified impact0/250 metrics describe measurable impact.
- Deployment scale10/10Rewards evidence of a live or scaled deployment and meaningful integration scope.
Evidence confidence88%
- Source authority20/25The main evidence is an official vendor-published customer story.
- Claim traceability23/2511 of 12 key fields link to a named source.
- Direct evidence20/204 direct source quotes retained.
- Date provenance8/10The publication date comes from publisher-controlled metadata.
- Source transparency9/102 source references document publisher, URL, authority, retrieval, and supported fields.
- Implementation specificity8/10Rewards source-backed components, data flow, quantified impact, and a specific outcome.
Traceability
Sources & provenance
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
- https://openai.com/index/khan-academy/OpenAIFirst-party
Khan Academy customer story
Supports- Customer
- Business problem
- Solution
- Products
- Solution components
- Data flow
- Integrations
- technical area
- use case category
- Business outcome
- outcome category
- Metrics
- Architecture
- evidence quotes
- Publication date
- https://builtin.com/company/khan-academyBuilt InIndependent secondary
Khan Academy Careers, Perks + Culture | Built In
The Organization JSON-LD lists addressLocality "Mountain View", addressCountry "US", and numberOfEmployees value 794.
Supports- Company size
Date evidence
- Publisher catalog date Mar 14, 2023 · OpenAI · First-partyOpen date source
Not separately sourced Industry, Region. Treat these normalized attributes as lower-confidence until a dedicated source is added.
Related
Similar use cases
OpenAI Sample Jun 11, 2026 University of Arizona University of Arizona researcher Chi-kwan Chan needed faster ways to create stable numerical approximations for black hole simulations because the equations are computationally intractable and new approaches can take… speed up these calculations by a factor of 1000 Codex OpenAI Sample Sep 22, 2025 SchoolAI SchoolAI needed to help teachers personalize instruction and spot struggling students earlier without letting AI simply hand answers to students or create unsafe, opaque classroom behavior. 1 million classrooms reached across 80+ countries Dot · GPT-4.1 · GPT-4o OpenAI Sample Sep 17, 2024 Arco Educação Arco Educação needed to reduce the time teachers spend on lesson planning and individualized education plans so they could focus more on direct instruction, especially for students with diverse learning needs. Arco serves over 3 million students OpenAI models · Teacher Assistant