Cognii is a conversational AI-based that provides personalized tutoring and assessment services for K-12 and higher education, helping students deeply understand concepts through natural language dialogue.

Cognii

Cognii’s core parameters and statistics

Cognii is a conversational AI tutoring and assessment platform for the K-12, higher education and corporate training markets. It is officially positioned as "artificial intelligence-driven scalable personalized education." Unlike Quizlet, Khan Academy and other platforms that focus on question banks and videos, Cognii's core interaction form is open question and answer + conversational tutoring - students answer open questions in their own language, and the AI ​​engine evaluates the quality of the answers at the semantic level and gives formative feedback in real time. This mechanism is closer to the effect of one-to-one tutoring by real people, rather than the traditional "question-answering-scoring" sequence.

Projects	Public Information
Official positioning	AI-based personalized education for K-12, higher education & corporate training
Core Products	Cognii Virtual Learning Assistant (VLA)
Target Markets	K-12 schools, universities, school districts, corporate training departments
Core technology stack	Natural language processing (grammatical analysis + deep semantics + conceptual level), machine learning evaluation engine
Assessment Accuracy	Study shows 96% agreement with human raters (short essay assessment)
Supported languages	English (undisclosed multi-language support plan)
Deployment method	Cloud SaaS, supports LMS embedding (LTI integration)
Company location	San Francisco, CA, USA
Established	Approximately 2013
Latest version	Unpublished (continuous iteration)
Number of registered users	Undisclosed

Assessment Accuracy: A controlled study cited on the official website showed that Cognii achieved 96% agreement with human raters in scoring short essays. This number comes from a controlled experiment. In actual classroom scenarios, scoring consistency may fluctuate due to subject coverage, question complexity, and data sparseness, but as an academic reference it is sufficient to illustrate the maturity of its natural language assessment engine.

Positioning boundaries: Cognii's positioning is "coaching and assessment" rather than "content creation". Teachers still need to write course questions and grading standards themselves, and the platform does not provide automatic generation of subject content. This is complementary to institutions that already have a complete course syllabus and question bank, but there is an obvious capability gap for teams that want to "automatically generate courses from scratch."

Cognii’s users and market recognition

Cognii's market recognition is mainly reflected in industry awards and academic partnerships. The number of public customers and revenue data are not disclosed.

Industry Awards: The official website publicly displays two authoritative awards——

• MassTLC Best of Education Technology Award (2016): Selected by the Massachusetts Technology Leadership Council (MassTLC) to recognize the most innovative education technology companies in New England.
• Reimagine Education Award for Best Innovation in Learning Assessment (2015): a global competition that received more than 500 project applications that year and was selected by an international panel of 40 judges.

Academic Cooperation: Cognii has established cooperation with universities and K-12 school districts. The official website mentions cooperation cases with institutions such as Drexel University, but does not disclose a detailed customer list or contract amount.

Market Positioning Comparison: In the AI ​​education tutoring track, Cognii has formed a differentiated competitive landscape with the following platforms:

Comparative Dimensions	Cognii	Khan Academy	Quizlet	Gradescope	Turnitin
Core interaction	Open dialogue + AI assessment	Video + question bank exercises	Flashcards + quizzes	Human scoring workflow	Paper plagiarism check
AI assessment capability	Semantic-level open question scoring	None (only multiple-choice/fill-in-the-blank questions to determine whether true or false)	None	Limited (group scoring assistance)	None (only duplicate checking)
Formative feedback	Real-time, personalized, iterable	Standard answer comparison	True or false feedback	Teacher manual required	None
1-to-1 coaching	Built-in conversational coaching	None	None	None	None
Pricing model	Customized by institution	Free (supported by donation)	Freemium	By student/course	By institution
Open Source	No	Yes (Partial)	No	No	No
Multi-language	English only	Multi-language support	Multi-language	English-based	Multi-language

As can be seen from the above table, Cognii is one of the few products currently available on the market in terms of the vertical capability of "AI-driven open question and answer assessment", but its brand awareness and user scale are far less than those of free platforms such as Khan Academy.

Cognii’s cost advantage: AI replaces manual process savings

Cognii's cost structure is not reflected in "absolutely low prices", but in the process savings brought about by "replacing labor with AI". Assessing its cost value requires dismantling it from three levels: the C-side, the institutional side, and the developer-side.

C-side/Individual Learners: Cognii does not sell directly to individual consumers. Students must obtain access through their school or educational institution. This means that individuals cannot purchase or trial independently, and the decision-making power rests entirely with the institution. This is not an option for individuals who want to improve their learning at their own expense.

B-side/Schools and Districts: The official standardized pricing table has not been disclosed, and the annual fee is customized according to the number of students or school size. Typical fee components include:

• Platform License: Billed based on the total number of students or active users, and the contract period is usually one or three academic years.
• Technical Support: Includes initial integration deployment and ongoing operation and maintenance support.
• Teacher Training: Some contracts include teacher usage training and best practice workshops.
• API Integration: If you need to deeply embed an existing LMS or a custom front end, additional API licensing fees may apply.

Corporate/Corporate Training: Solutions for corporate learning and development (L&D) departments are also quoted on an on-demand basis and usually include integration adaptation with internal LMS (such as Cornerstone, SAP SuccessFactors), exclusive SLA and data isolation terms.

Quantitative deduction of cost reduction and efficiency improvement (The following is an estimation based on public information, unofficial data):

Cost items	Traditional manual mode (estimate)	Cognii assisted mode (deduction)	Savings margin (deduction)
Grading hours per week (200 students)	Teacher 10-15 hours	AI initial assessment + teacher review 3-5 hours	60-70%
Course tutoring response latency	24-48 hours (asynchronous)	Real-time (<1 second)	Not comparable
Frequency of formative assessment	3-5 times per semester (manpower limit)	Multiple times per week	5-10x improvement
Teacher training investment	2-3 days of intensive training per semester	First deployment + online guidance	About 50%

Hidden costs: Cognii adoption requires institutions to invest in initial integration time (LTI docking, data migration, teacher training), which is easily underestimated in purchasing decisions. In addition, AI scoring may encounter "algorithmic fairness" doubts among on-campus review committees, requiring additional explanatory documentation and pilot data to build trust.

Position-level cost reduction and efficiency improvement mapping (deduction): There are significant differences in the impact of Cognii on the working hours of different education positions——

• Teachers: Weekly grading hours can be reduced from 10-15 hours to 3-5 hours (saving 60-70%), and the freed time can be used for small class tutoring and instructional design instead of mechanical correction.
• Teaching Assistant/TA: In a formative assessment scenario, TA's grading workload can be reduced by 80-90%, but needs to shift to the new responsibility of "AI grading quality sampling".
• Academic Administrator: Data analysis and report generation time has been reduced from 2-3 hours per week (manual compilation) to almost zero (automated dashboard), but additional time needs to be invested to interpret the data and develop intervention strategies.
• IT Support Team: The initial integration deployment requires 20-40 man-days of investment (LTI configuration, data migration, permission settings), and then the daily operation and maintenance burden is low (<2 man-days/month).

Cognii’s main features

Cognii's product capabilities revolve around the core axis of "open question and answer assessment + conversational coaching", and each function forms a complete relationship from assessment to feedback to path adjustment.

• Open Response Assessment: Students write answers in natural language, and the AI ​​engine evaluates the completeness, accuracy, and logical structure of the content at the semantic level, rather than just checking keyword matching or sentence fluency. Supports a variety of question types from short questions and answers to short essays. Value: Breaking through the limitations of traditional multiple-choice questions that cannot measure higher-order thinking abilities, allowing assessment to return to "understanding" itself.

• Formative Assessment & Immediate Feedback: Students receive scores and personalized feedback immediately after submitting their answers, which clearly points out the correct and missing concept points in the answer instead of just giving a right/wrong judgment. Students can revise and resubmit multiple times and receive iterative feedback each time until they master it. Value: Transforming assessment from an "end-point test" to a "part of the learning process" is consistent with the Bloom 2-sigma tutoring effect - that is, one-on-one tutoring can improve student performance by two standard deviations.

• Conversational Tutoring: Cognii calls itself "Siri in education" and supports students to conduct multiple rounds of question and answer dialogues with AI. When AI discovers conceptual misunderstandings, it will proactively ask questions and guide reflection, rather than directly giving standard answers. This mechanism simulates the Socratic teaching method of a real instructor. Value: It solves the classic contradiction in the education field that "one-on-one personalized tutoring cannot be scaled up".

• Adaptive Personalization: Based on the semantic evaluation results of each open-ended answer, the system dynamically adjusts the difficulty, topic and depth of tutoring for subsequent questions. Students' knowledge graph is gradually constructed, and weak points are automatically marked and targeted exercises are arranged. Value: Allow each student to obtain a learning pace that matches their current level and avoid "one size fits all" teaching progress.

• High-Resolution Analytics: Teachers and administrators can view the heat map of the class's overall knowledge point mastery, each student's concept mastery trajectory, the distribution of common error patterns, and the rate of learning progress. The granularity of analysis can be as deep as "a student's level of understanding on a certain knowledge point in a certain subject." Value: Shifting teaching decision-making from "experience-driven" to "data-driven" allows teachers to accurately locate students and weaknesses that need intervention.

• LMS and course content integration: Supports LTI standard integration and can be connected with mainstream learning management systems such as Canvas, Blackboard, and Moodle. Teachers can embed Cognii assessment activities directly within the framework of existing courses without having to switch systems. It also supports custom question banks and scoring standards. Value: Reduce technology migration resistance and implement AI enhancements within teachers’ existing workflows.

Cognii’s model and version evolution

Cognii has not disclosed the detailed internal model version number and release timeline, but its technical iterations can be sorted out from the official website information and industry public milestones.

Early exploration (~2013 — 2015)

Cognii was founded in San Francisco, and the founding team started at the intersection of natural language processing and educational technology. Won the Reimagine Education Best Learning Assessment Innovation Award in 2015, indicating that its early open-response assessment technology has been recognized in academic reviews. The product form at this stage is mainly "automatic scoring engine" and has not yet formed a complete conversational tutoring experience.

Product shaping and commercialization (2016 — 2020)

Winning the MassTLC Award for Best Educational Technology in 2016 marked the product’s transition from research prototype to commercially deployable solution. During this period, Cognii gradually launched the Virtual Learning Assistant (VLA) product line, upgrading the assessment engine into a tutoring system with conversational interaction capabilities. Examples of cooperation with universities such as Drexel University were established during this period. The addition of LTI integration capabilities enables Cognii to be embedded into mainstream LMS workflows.

Continuous iteration period (2021 - present)

Cognii has not released a specific version number or release instructions, but the official product page continues to display the following technology evolution directions:

• Syntax Analysis + Deep Semantics + Concept Level: Upgraded from the initial keyword matching to a multi-layer NLP parsing architecture.
• Improved Assessment Accuracy: Gradually improved from early "acceptable" levels to 96% human consistency shown in studies.
• Adaptive path optimization: The personalized recommendation engine transitions from rule-driven to data-driven.

The latest version information has not been made public. There is no changelog or release notes page on the official website. The product iteration rhythm belongs to the "continuous delivery" model rather than the "major version release" model. Organizations should confirm the currently deployed engine version and feature set directly with the sales team when evaluating.

Summary of key milestones:

Time	Event	Meaning
~2013	Cognii was established, located in San Francisco	The company was founded and started NLP+EdTech
2015	Won the Reimagine Education Best Learning Assessment Innovation Award	Technical solution recognized by international review
2016	Won the MassTLC Best Educational Technology Award	Product commercialization capabilities recognized by regional authorities
~2016-2018	Launch of Virtual Learning Assistant	Upgrade from assessment engine to conversational tutoring
~2020	LTI integration + LMS docking	Enter the school procurement standard workflow
So far	Continuously iterating NLP engine and adaptive path	Undisclosed version number, delivered on demand

Cognii’s technical advantages

Cognii's technology route selection revolves around a core contradiction: the balance between the evaluation accuracy of open question and answer and the cost of large-scale deployment. Its NLP engine is not a general-purpose large language model (such as the GPT series), but a vertical model designed specifically for educational assessment scenarios.

Three-layer semantic parsing architecture: Cognii’s evaluation engine contains three levels from bottom to top——

1. Language Syntax: Analyze how words are combined into phrases, sentences and paragraphs, and identify basic grammatical correctness and structural integrity.
2. Deeper Semantics: Understand the meaning of words, phrases and sentences in the context of the question, rather than literal matching. This layer is the essential difference between Cognii and traditional keyword scoring systems.
3. Concept Hierarchy: Determine which semantic concepts are included in students’ answers, which concepts are missing, and which concepts are misunderstood. This is the core of realizing the ability to "identify knowledge blind spots".

Evaluation accuracy vs. general AI: Cognii chooses a proprietary NLP model for vertical education scenarios instead of directly calling general-purpose large models such as GPT. The advantages of this technical route are: the scoring standards are controllable (do not drift with the general model version), the reasoning cost is predictable, and the student data does not leave the domain. The price is: lack of zero-sample adaptability of the general model, and teachers need to pre-set scoring standards or provide reference answers for each question.

Technical path comparison with general LLM:

Comparative dimensions	Cognii proprietary NLP engine	Generic LLM (such as GPT-4) for scoring
Evaluation method	Syntax + semantics + concept hierarchical structured analysis	End-to-end Prompt reasoning
Scoring consistency	Structured rules are reproducible	Affected by temperature coefficient and Prompt wording
Data privacy	Localized deployment/cloud controllable	Please pay attention to the API data usage terms
Teacher customized	Preset grading standards/reference answers required	Zero-shot can be assessed, but the quality is unstable
Cost structure	Based on annual contract fee, no Token billing	Based on Token consumption, large-scale use costs increase linearly
Suitable for question types	Short answers, short papers (50-500 words)	No limit on question types, but long texts are expensive

Causal explanation: why 96% consistency can be achieved - The three-layer parsing architecture allows Cognii not only to determine "whether the answer is correct", but also to locate the precise problem of "concept A is correct but concept B is wrong". This fine-grained evaluation capability is not possible with traditional text matching or simple ML classifiers. The 96% consistency shown in the study is more reliable on high-frequency standard questions, but it may fall back to the 85-90% level on creative writing or highly open-ended questions, which requires teacher review.

Engineering implementation of adaptive mechanism: Cognii's adaptive path system builds student knowledge graphs based on the results of concept level analysis. After each assessment, the system updates the "mastery" indicator in the map and determines the direction of the next question through preset course structure rules (rather than a pure neural network). This approach is more interpretable than a purely data-driven approach, but the rule maintenance cost increases with the number of courses.

How to use Cognii

Cognii does not provide a standalone consumer application (no App Store personal version), and all access is through educational institutions.

How to use	Applicable objects	Activation prerequisites	Main operations
LMS Embedded (LTI)	Schools with Canvas/Blackboard/Moodle deployed	Institution purchases license + IT completes LTI configuration	Teachers add Cognii activities to courses and students answer directly within the LMS
Cognii independent portal	Institutions that do not use mainstream LMS	Institutions purchase licenses + academic administrators open classes	Students log in to the Cognii platform and complete exercises by class/course
API integration	Online education platform MOOC, e-book provider	Business agreement + technology docking	Embed Cognii assessment engine into own products

Typical Teacher Workflow (single assessment activity):

1. Create a new Cognii assessment activity in your LMS, select or write open-ended questions, and set scoring criteria.
2. Set the number of attempts allowed (3-5 is recommended to give full play to the formative assessment effect) and the opening time window.
3. After students answer, AI will score in real time and give personalized feedback. Students can review feedback and revise resubmissions.
4. Teachers log in to the dashboard to view the heat map of class mastery, distribution of common errors, and disputed answers that require manual review.
5. The teacher conducts random inspections and reviews of the AI ​​scores according to the human-machine collaboration strategy (see "Human-machine collaboration boundary" for the review ratio).

Typical Student Experience:

1. Receive open-ended questions (such as "Explain the energy conversion process of photosynthesis").
2. Enter your own answer (50-200 words of natural language).
3. Get an immediate rating + concept-level feedback: "You described the light reaction stages correctly, but the carbon fixation process in the dark reaction (Calvin cycle) is incomplete."
4. Revise your answer based on feedback and submit again.
5. After 2-3 rounds of iteration, gradually improve under the guidance of AI until you master the core concepts.

Human-machine collaboration boundary: The use of Cognii requires a clear definition of AI automation and manual confirmation points——

• Can be 100% automated: Initial evaluation and instant feedback of standard open-ended questions and answers, multiple rounds of iterative tutoring for students, statistics of common error patterns and heat map generation of knowledge point mastery, and simulated scoring of standardized exams.
• Human-in-the-loop required: final scores of high-stakes exams (final exams, entrance exams), student answer reviews involving controversial or sensitive topics, answers with AI score confidence lower than the threshold (such as <80%), teachers' initial setting and periodic adjustment of scoring standards for assessment activities, and the formulation and implementation of class intervention strategies.
• Recommended human review ratio: 100% teacher review recommended for the first batch of assessment activities to establish a trust baseline; targeted review triggered by 10-20% random sampling + confidence threshold during the regular run phase; revert to 50% review for 2-3 weeks after each change in course content or grading criteria to calibrate consistency.

Integration Note: LTI 1.3 standard compatibility needs to be confirmed before purchasing; some older LMS versions may not support all Cognii features. In the API integration scenario, you need to pay attention to the concurrency limit and data reflow delay. It is recommended to clarify the SLA terms in the contract.

Cognii product pricing

Cognii does not disclose a standardized pricing list, and all plans adopt a business model customized according to the size of the organization. The following information is based on the description on the official website and the pricing logic of similar products in the industry. The official quotation shall prevail.

C-side/Individual Learners: No direct sales. Students cannot purchase it independently and must access it through their school or training institution.

B-side/Schools and Districts: Pricing typically includes the following fee items—

Cost structure	Description	Typical range (deduction)
Platform license fee	Billed based on the total number of students or active users	Undisclosed, refer to similar AI evaluation platforms about $5-15/student/year
Initial deployment fee	LTI integration, data migration, teacher training	One-time, depending on organization size
Annual technical support	Remote support + online resources	Included in license fee or quoted separately
Teacher training workshops	Optional value-added services	Quotation by session
API integration	Additional integration requirements	Billed by developer day or included in annual fee

Enterprise/Corporate Training: Plans for enterprise L&D departments typically upgrade the education version with data isolation, dedicated SLAs, and advanced analytics. The price is significantly higher than the education version.

Free Trial: The official website does not provide a self-service registration trial entrance. Organizations need to apply for demonstration or trial through the official website contact form or email ([email protected]), and the process is "Business Communication -> Technical Verification -> Small-Scale Pilot -> Formal Signing".

Focus on contract terms:

• Data Privacy: Written confirmation is required that student data (FERPA, COPPA compliant) will not be used for model secondary training.
• Contract Duration: Typically 1-3 years, discounts may be available on longer term contracts but be aware of exit clauses.
• Service Level: The SLA should include availability commitments (recommended ≥99.5%) and support response times.
• Data Export: Whether the student data can be completely exported after the contract expires should be confirmed in advance.

Cognii application scenarios

Cognii's implementation scenarios focus on educational issues that "require high-frequency formative assessment + lack sufficient manpower for one-to-one personalized tutoring".

• K-12 subject formative assessment: In subjects such as science, mathematics, and English that require open thinking, teachers assign 2-3 open-ended questions every week. After students answer, AI will score them immediately and guide corrections. Teachers save 5-8 hours of grading time per week and can turn their energy to targeted small class tutoring. Key points of verification: Consistent sampling of AI scoring and teacher scoring within the class (it is recommended to sample 5-10 manual reviews for each subject in each class).

• Preliminary evaluation of university humanities and social science papers: In courses that require exposition skills such as history, philosophy, and psychology, AI first provides preliminary scoring and conceptual-level feedback to students' short papers. After students revise, they submit the final version for final review by the teacher. Teacher feedback time is shortened from 3-5 working days to a few hours, and students' willingness to revise is significantly improved by receiving immediate feedback. Key points of verification: The boundaries of AI's ability to evaluate creativity and argumentation logic - it is currently more suitable for factual discussions and has limited ability to evaluate highly personal opinions.

• Distance Education and MOOC Tutor Replacement: Embed Cognii assessment activities in online courses to provide personalized feedback to thousands of students at the same time, making up for the response delay of asynchronous communication between teachers and students in remote scenarios. Implementation Tips: In API integration mode, the concurrency limit and response time need to be evaluated in advance. Official concurrency capacity data has not been disclosed, and it is recommended to complete a stress test before signing the contract.

• Corporate Training Knowledge Verification: Embed open-ended Q&A tests in internal L&D courses. After employees complete the course, AI will automatically assess their understanding, and HR can view a heat map of the team's overall mastery. Value: More accurately identifies "coping-based learning" than traditional multiple-choice questions - employees must explain knowledge points in their own words and cannot pass by guessing.

• Standardized test preparation: In response to the increasing number of standardized tests that use open question types (short answer), students can use Cognii to conduct targeted exercises. AI simulates the scoring standards and promptly points out loopholes in answer ideas. Value: Upgrade your exam preparation from "study multiple choice questions" to "constructive learning".

Applicable groups of Cognii

Cognii's service roles cover three types of core users in education scenarios, and the usage boundaries of each type of role are different.

• Teachers and Instructional Designers: Benefit from the time saved by AI automatic grading and fine-grained analysis of data. Adaptation Conditions: Willing to invest 2-4 hours in setting up learning assessment activities and setting scoring standards; subjects mainly focus on factual knowledge (science, mathematics, language, history, etc.); ROI is most obvious when the class size is more than 30 people. Not suitable: Purely creative writing (poetry, personal narrative) teachers, AI has limited ability to evaluate creative content, and may limit students' freedom of expression due to over-standardized feedback.

• Students and Learners: Benefit from instant feedback and unlimited iterations of the learning experience. Adaptation Conditions: Have basic written expression skills (English); students with strong self-motivation can benefit the most from multiple iterations; suitable for scenarios that require building deep understanding rather than short-term memory. Not suitable: Preschool/young children (requires picture-text interaction, Cognii is text-based); students who are not good at written expression may be frustrated by AI's text evaluation.

• Education Administrators and IT Teams: Benefit from a unified assessment data dashboard that integrates seamlessly with your LMS. Adaptation Conditions: The institution has deployed mainstream LMS such as Canvas/Blackboard/Moodle; the IT team has the ability to complete LTI integration configuration; managers are willing to delegate some grading responsibilities to AI. Not suitable: For organizations that have not standardized LMS, integration costs will increase significantly; managers who hold "zero error" expectations for AI scoring need to first establish reasonable expectations through small-scale pilots.

• Corporate Training and HR Departments: Benefit from automated knowledge verification and skills assessment. Adaptation conditions: The training content is mainly based on the standard knowledge system (compliance training, product knowledge, process operations, etc.). Not suitable: Soft skills training (leadership, communication skills, etc.), AI text assessment is difficult to measure abilities at the behavioral level.

Summary and Outlook of Cognii

Cognii has a clear technical moat in the vertical track of "conversational AI educational assessment" - its proprietary three-layer NLP assessment engine and 96% human consistency research results are academically verified core assets. It is also one of the few products on the market currently that has truly commercialized "open question and answer automatic scoring + formative feedback".

Core Competencies: Essentially different from multiple-choice platforms (Quizlet, Khan Academy) and manual grading tools (Gradescope), Cognii uses AI to realize the "reading-understanding-evaluation-guidance" process that only human tutors can provide in the past. For institutions that already have course content but lack personalized tutoring manpower, this is a cost-controllable upgrade path.

Current limitations and uncertainties: Only supports English, no clear multilingual roadmap, and is not applicable to K-12 or ESL scenarios in non-English speaking countries; it is not open source and has no community version, so institutions cannot expand or customize it on their own; pricing is not transparent, and the communication cost from contact to obtaining a quote is high for small and medium-sized institutions; there is no independent consumer version, and individual learners cannot use it directly; the version iteration process is not public, making it difficult for institutions to evaluate the predictability of technology evolution.

Competitive landscape observation: With the rapid improvement of the scoring capabilities of general large models such as GPT-4, Cognii faces the long-term risk of "general AI dimensionality reduction attack" - if schools can directly use GPT to complete open question and answer assessments at a low enough cost, the differentiated value of the proprietary NLP engine will be compressed. However, in the short term, the issues of scoring consistency, data privacy and token cost of general models still retain a window period for Cognii.

Procurement/Adoption Risk Assessment: It is recommended that institutions advance in "three phases" - the first phase (1-2 months), apply for demonstration and complete the Pilot within a small range of 1-2 classes, focusing on verifying the consistency of AI scoring and teacher scoring, student acceptance and IT integration stability; the second phase (1 semester), expand the Pilot to 5-10 classes, and establish a teacher review and random inspection mechanism, collect test score comparison data to evaluate "classes using Cognii vs. The difference in learning effects of traditional classes; in the third stage (1-2 years), decisions are made on whether to expand to the whole district/school based on the support of quantitative data. Before signing the contract, it must be confirmed in writing that student data will not be used for model secondary training. LTI 1.3 compatibility has been fully tested in the IT environment, as well as the data export format and path after the contract expires.

Related tools: khanmigo, quizlet

Version Info

• Cognii latest version ：There is no official precise date yet, and the dialogue coaching engine will continue to be iterated. （~2026）
• Cognii V1 ：There is no official precise date yet, but early versions of Cognii will launch conversational learning assessment capabilities. （~2016）