Learnosity AI

Positioning

Learnosity is not a product you log into — it is assessment infrastructure you embed. The company sells a set of APIs (Author, Questions, Math, Assessments, Analytics) that publishers, edtech platforms and assessment providers build into their own products, and the current pitch on learnosity.com is blunt about the job: ship assessment products in weeks, not quarters, and let Learnosity handle the plumbing. For 2026 the headline addition is agentic AI grading, joining earlier AI-assisted item generation. The name in our catalog, Learnosity AI, slightly misrepresents the scope: AI is a feature layer on a mature assessment engine, not the product itself. Anyone evaluating this needs to be a builder with an engineering team, not a teacher shopping for a quiz tool.

What it actually does

The product surface, as documented on learnosity.com and its pricing page, splits into five API families. Author is the item authoring environment where content teams write questions. Questions is the rendering layer — the site claims 50+ question types, from multiple choice through the drag-and-drop and graphing formats seen in the interactive demos. Math handles equation input and rendering. Assessments covers delivery and scoring, split into Essentials (included in all tiers) and Premium (paid add-on below the top tier). Analytics covers reporting, also Essentials and Premium. On top of these, the current site advertises agentic AI grading and AI-assisted item generation, and a live progress-reporting demo showing per-student status and session scores.

The commercial structure is tiered by user volume: Standard (up to 100,000 users), Scale (100,000–500,000) and Enterprise (500,000+). All three include Author, Questions, Math, Assessments Essentials and Analytics Essentials. Below Enterprise, Assessments Premium and Analytics Premium cost extra; Enterprise includes them. Tenancy and residency options also step up with tier — single tenancy is not available on Standard, is a paid add-on on Scale, and is included at Enterprise. Support tiers map Silver/Gold/Platinum to the three plans with SLA uptime guarantees of 99.8%, 99.9% and 99.9%, plus 24/7/365 product, developer and author support staffed across New York, Dublin and Sydney. Solution consultancy runs 6 hours per year on Standard, 12 on Scale, bespoke on Enterprise.

Pricing itself is quote-based. Nothing is published on learnosity.com/pricing beyond the tier structure. A third-party directory (softwarefinder.com) puts a base plan at $50,000/year; I could not verify that with the vendor, so I treat it as an estimate only. The batch data I was given suggested a $500 floor, which I could not corroborate anywhere and consider wrong.

My testing notes

When I worked through the item authoring demo embedded on the homepage, the ratio-based flour-and-sugar question showed the detail level of the question engine: drag-and-drop tokens with validation states (140 g placed, 490 g placed) and distractors, plus a live progress report showing Alex Chen at 86/100 with a class-mean comparison. What I look for in this category is exactly that — the plumbing of scoring and reporting visible in the demo rather than promised in prose. The demo is clearly a curated example, but it reflects capabilities (partial credit on placed tokens, item-level analytics) that are hard to fake.

The second thing I checked was the security and compliance surface, because assessment data of minors is the hardest privacy problem in edtech. The site lists ISO 27001 certification, role-based access, encryption at rest and in transit, regular third-party penetration testing, and GDPR/FERPA/COPPA compliance with regional data residency. For an API vendor selling into districts and publishers, that list is the right shape, and the FERPA/COPPA posture is the single most important spec here — most consumer AI study tools have nothing comparable.

Third, the customer roster on the homepage is the strongest reliability signal I found: Pearson, HMH, College Board, Wiley, PowerSchool, Instructure, iCEV, Cuemath and BARBRI are all named as customers (vendor-published logos). Companies like Pearson and College Board do not embed an assessment API without serious procurement scrutiny. The site also publishes an NPS of 67+ and 47.4M+ learners served — vendor claims, but the logo wall makes them plausible. Learnosity also states >99.97% delivered uptime per API against SLA promises of 99.8–99.9%.

What implementation actually looks like

Teams that buy Learnosity buy a project, and my advice to anyone scoping it is to budget for the integration honestly. The typical implementation runs in three stages: an item migration or authoring bootstrap (moving existing question banks into the Author API or rebuilding them in its format), an embedding pass (rendering Questions and Assessments inside your product's UI and wiring scoring and reporting endpoints), and an analytics integration (feeding your warehouse or dashboard from the Analytics API). Learnosity's own pitch is 'weeks, not quarters', and for a team with existing content and a clear product shell that is plausible — the API surface covers authoring, delivery, scoring and reporting, so the plumbing is genuinely pre-built. The parts that never compress are content quality and workflow design: authoring teams need training on item types and metadata conventions, and your product team needs to decide how adaptive behaviour and analytics surfaces appear to learners.

Third-party review aggregators give a sanity check on vendor claims: Capterra and G2 both list Learnosity with reviews from the edtech-builder audience rather than end users, which is consistent with what it is. The absence of end-teacher reviews is not a red flag here; it is the correct signal for an API product.

The consultancy allowance deserves a line in your scoping even though it is included: six hours per year on Standard, twelve on Scale. In my experience those hours are where you learn the conventions that prevent rework, and organisations below the Scale tier should expect to either self-serve from documentation or buy partner help. The 24/7 developer support across New York, Dublin and Sydney is the safety net that makes the small consultancy allowance tolerable — support answers the how, consultancy answers the should.

What it does not do

There is no self-serve signup and no free tier — you cannot evaluate this without a sales conversation and a contract, which is normal for the category but means the learning curve starts at procurement. It is not an LMS and not a course platform: there is no learner-facing app, no content library, no curriculum. AI grading and item generation are new (the site brands 2026 as "now with agentic AI grading"), and I found no published accuracy benchmarks, human-in-the-loop controls documentation or grading-fidelity studies — for high-stakes assessment, that evidence gap matters and I would demand it in a trial. The AI features also appear to track the Premium tiering, so the advertised AI may cost extra below Enterprise. Implementation requires engineering: API integration, item migration, authoring workflow design. And "50+ question types" is a vendor count; complex custom interactions still need custom development.

Verdict

Learnosity is the most credible assessment infrastructure I reviewed in this batch, and for a publisher or edtech platform the buy-versus-build maths is not close — the pricing page itself claims in-house development costs fall 50% or more, and while that is a vendor figure, the direction is right. It loses points for opacity: no published pricing, unverifiable entry cost, AI features with no published accuracy evidence, and premium gating on the analytics that make the data valuable. I rate it 4.1: the rubric's 40% weight on core functionality depth reflects a genuinely deep, decade-mature assessment engine with real enterprise customers, and the value, reliability and support scores follow the SLA, compliance and roster evidence; pricing opacity and the early state of the AI grading keep it out of the top band.

Sources

• https://learnosity.com/
• https://learnosity.com/pricing
• https://learnosity.com/build
• https://aws.amazon.com/marketplace/pp/prodview-veaxd4xh7fyhs
• https://softwarefinder.com/lms/learnosity
• https://learnosity.com/contact-sales
• https://www.capterra.com/p/171038/Learnosity
• https://www.g2.com/products/learnosity/reviews
• https://www.capterra.co.uk/software/171038/learnosity
• https://aws.amazon.com/marketplace/pp/prodview-njr7giu5mhpq4