The six AI writing feedback tools worth a teacher's time are Brisk Teaching, Class Companion, EnlightenAI, Olex.AI, Grade Direct and Grade It AI. Brisk Teaching is the best free start because it works inside the Google Doc a student already has open. Five of the six now publish a statement about training AI on student work, each worded differently; Class Companion's privacy policy does not.
Last checked November 16, 2026. Every price, free tier and data statement below was read on the vendor's own page that day.
Which AI writing feedback tools should teachers use, and what does each cost?
| Tool | Best for | Free tier | Paid price, checked November 16, 2026 | Student work and AI training, in the vendor's words | Where it is stored |
|---|---|---|---|---|---|
| Brisk Teaching | Feedback inside the document a student is already writing in | Yes: $0 "For Teachers"; batch feedback has an upload limit, figure not printed (plans) | Premium and Intelligence: "Custom Pricing" (plans) | "Doesn't use student or teacher data to train or fine-tune AI models" (privacy center) | Amazon Web Services, region not named (security FAQs) |
| Class Companion | Feedback a student acts on mid-draft | Yes: "Use the product for free as a teacher" (plans) | School and district tiers: no public pricing (plans) | No training statement; the policy says de-identified data is used "for research and development purposes" (privacy) | Servers in the United States (privacy) |
| EnlightenAI | Applying your own rubric once it has learned how you score | Yes: "Free for teachers" (essay grader) | No public pricing; schools and districts talk to the team | "No AI provider we use has the right to train models on any data processed on EnlightenAI's behalf" (privacy) | A third-party host, location not named; our directory record says United Kingdom |
| Olex.AI | A class set of thirty essays in one pass, handwriting included | None stated; a "bespoke trial" on request (FAQs) | Quote based on school size (FAQs) | "Student work is never used outside of Olex to train any global AI models" (FAQs) | Not stated by the vendor; our directory record says United Kingdom |
| Grade Direct | Written responses in any subject at a price you can see | Yes: $0, 150 credits at signup, then up to 50 free a month (pricing) | Teacher Pro $25 for 6,000 credits, no expiration (pricing) | "We do not use Pupil Data to train AI models" (privacy) | DigitalOcean, United States with EU data center options (privacy) |
| Grade It AI | Grading English essays against a named exam board's mark scheme | No; a one-essay demo with no sign-up (site) | Starter $129 a year for 30 gradings a month; Pro $249 a year for 80 (pricing) | Does "not use your content to train models of our own"; essays are processed by OpenAI and Google (privacy) | Amazon Web Services, region not named (privacy) |
Our verdict. Start with Brisk Teaching if your students draft in Google Docs and you have no budget. Choose Class Companion if the feedback should reach the student during the task, EnlightenAI if your department grades to a shared rubric, and Olex.AI if you grade stacks of handwritten English papers in the UK system. Grade Direct and Grade It AI are the two with a dollar price you can budget against without a sales call. That is our editorial view from the vendors' published pages; nobody here graded a class set with them.
The fastest way to give written feedback on student essays has less to do with the tool than with the checkpoint. Run one rubric pass over the whole class set, read the feedback instead of thirty essays, and you learn the three or four criteria the class is failing together. That is one mini-lesson instead of thirty sets of comments.
How do you pick the right feedback tool for your class?
Work down these four questions in order. The first one usually settles it.
Where does the writing already live? If your students draft in Google Docs, a tool that reads those files directly will get used, and a tool that needs uploading will not. Brisk Teaching runs inside the browser; EnlightenAI pulls a class set from Google Classroom or Canvas; Olex.AI, Grade Direct and Grade It AI take photos or uploads.
Who reads the feedback first, the student or you? Two different products hide under one label. Class Companion and Brisk Boost, Brisk's student-facing side, put comments in front of the student while the draft is still warm. Olex.AI and Grade It AI put a graded set in front of you, and nothing reaches a student until you release it. Pick the one that matches what you want to happen.
Does the tool learn your rubric, or bring its own? EnlightenAI is built around calibration: you grade five submissions with it and it learns how you score each rubric row. Most others accept a rubric and weight it their own way. If your department has agreed criteria, calibration is worth the setup.
What happens to the writing afterward? This answer changed since October. Five of the six now publish a statement about training AI models on student work, and the statements differ in scope: Brisk Teaching's covers student and teacher data, Grade Direct's covers "Pupil Data," Olex.AI's covers training "outside of Olex," EnlightenAI's covers its AI providers, and Grade It AI's covers only "models of our own." Read the exact sentence in the table before you rely on it, and take it to whoever signs data agreements in your district.
Browse the writing, feedback and academic integrity category to see the full set alongside these six. On November 16, 2026 we counted the category from the directory's own data: 50 of the 153 listed tools carry that tag, of which 25 are freemium, 6 free, 13 paid and 6 quote-only. Nine of the 50 have a 9ine privacy and security assessment on file, and 11 show no data-hosting location at all. (Method: one read of the directory's full tool list, counted on each record's category, pricing-type, assessment and hosting fields.)
What changed on these vendors' pages between October and November 2026?
We reread every page this article cites. Seven things had moved or did not match, and each one changes what you should ask the vendor.
- Four more vendors now state a position on AI training. In October this page said Olex.AI was the only one of the six to do so. Brisk Teaching, EnlightenAI, Grade Direct and Grade It AI each publish a sentence now, quoted in the table above.
- Olex.AI's sentence is narrower than the one we quoted in October. Its FAQ now reads: "Student work is never used outside of Olex to train any global AI models, ensuring all data remains private to the school" (FAQs). "Outside of Olex" and "global" are the words to ask about.
- Grade Direct has left our directory. The listing under its former name, Graded Pro, no longer resolves, so this page now links to the vendor.
- Grade Direct's two price claims do not agree. The page says it "costs less than $10 per month" for teachers and also that "weekly grading for 125 students (e.g. 5 classes) requires around 4,500 credits per month" (pricing). At $25 for 6,000 credits, our arithmetic puts 4,500 credits at $18.75 a month. Under $10 a month buys about 2,400 credits.
- Grade It AI shows two sets of prices. The pricing page shows Starter at $10.75 a month billed yearly as $129; the homepage FAQ says "Starter is $12.99 per month for 30 gradings, and Pro is $24.99 for 80" (site). The higher pair is probably the month-to-month rate behind the page's Monthly toggle, which we could not open: unverified.
- Brisk Teaching describes its security audit two ways. The privacy center says "SOC 2 Type II compliant" (privacy center); the plans page FAQ says Brisk is "working toward SOC 2 Type II" (plans). Ask which is current.
- Two vendors' privacy policies raise a dating question. Class Companion's is marked "Last Updated: 29th April, 2024" (privacy). EnlightenAI serves two copies with different dates, 08/07/2026 at one address and 10/23/2026 at another.
Brisk Teaching: the free option that meets students where they write
Brisk Teaching is the best fit for a teacher who wants to start on Monday without a purchase order, because it is a Chrome and Edge extension that works on documents that already exist. The vendor says it works inside Canvas, Schoology and Google Classroom, "as well as Google Docs, Slides, YouTube, Microsoft tools, PDFs, and any webpage" (plans). It carries a 4.77 rating from 35 reviews in our directory, the largest review count of any tool in this comparison.
The plans table shows a check mark in all three plan columns for glow and grow, next steps and rubric criteria feedback, and for Brisk Boost, the student-facing side. The targeted feedback generator shows two check marks, which we read as the two paid plans; the check marks are images with no column labels in the page text, so confirm it on the page.
Watch out for: it lives in a browser extension, so a school on locked-down devices needs its IT team to allow it first. "Unlimited usage" also shows two check marks rather than three, and the batch feedback row reads "Upload limit" for one plan and "No limit" for the other two. We read the limited plan as Free, and the page prints no figure for the limit: unverified. The paid tiers carry no public price.
Pricing: Free "For Teachers" $0; Premium "For Schools & Districts" and Intelligence "For Districts" are both "Custom Pricing" (plans), checked November 16, 2026.
Safety: the vendor states that student and teacher inputs "are never used to train, fine-tune, or improve AI models" (privacy center) and that it "does not store the underlying student work it processes (e.g., the content of a student essay)" (security FAQs). Data sits in Amazon Web Services databases, with no region named on the pages we read. The privacy notice, effective September 7, 2026, says Brisk "may choose to make Brisk Boost available to children under the age of 13," with the school providing consent (privacy notice). The vendor says it has signed the National Data Privacy Agreement with a list of states dated November 2025 and is "open to signing district-level agreements" elsewhere. It also calls itself "fully compliant with FERPA, COPPA, and GDPR," which is the vendor's statement, not a finding: FERPA duties sit with your district, so check whether your state is on that list. Our directory record still shows hosting and age as unknown, and holds a 9ine assessment for it.
Class Companion: feedback a student gets while the draft is still warm
Class Companion is the best fit for the teacher whose complaint is that feedback arrives too late to matter, because its free teacher tier lists "Instant AI Feedback" and "Student Progress Insights," which put a response in front of the student during the task (plans). The free tier also opens a library the vendor describes as "1,000+ assignments and rubrics built by Class Companion and teachers."
Watch out for: nothing at school scale has a public price. Our directory holds a single review for it, rated 4.00, so treat the rating as thin.
Pricing: free for teachers, with no limit stated; school and district tiers show no public pricing (plans), checked November 16, 2026.
Safety: this is the one tool of the six with no sentence about AI training. Its privacy policy, last updated April 29, 2024, says "We do not share teacher or student personally identifiable information with AI service providers," then adds that it de-identifies data and uses it "for research and development purposes" and "may retain de-identified data after you delete your account" (privacy). Servers are in the United States, and the policy says "we require parental consent for students under the age of 13." The plans page lists "FERPA Compliant" on the free tier and "District DPA & SSO" on the school tier; we could not find, in the text we read on November 16, the two longer compliance sentences this article quoted in October. The policy's own wording is more careful: "We support schools in complying with FERPA, and we comply with COPPA." Ask whether the district agreement is available to you before a class uses the free tier.
EnlightenAI: the one that learns how you score your own rubric
EnlightenAI is the best fit for a department that has agreed criteria and wants the machine to apply them, because, in the vendor's words, it "learns how you score, drafts a grade and feedback for every student, and you approve each one" (essay grader). You "train the AI by grading five submissions together" (home), then pull a class set from Google Classroom or Canvas, or upload Google Docs and Word files. The same page offers official state and AP rubrics for 40 states and 6 AP exams.
The vendor reports one accuracy comparison: on 147 grade 8 essays from DREAM Charter Schools, its scores matched the teachers' exactly 70 percent of the time and fell within one point for the other 30 percent, a count that "includes the five essays used to train EnlightenAI" (essay grader). That is the provider's own figure, not independent research.
Watch out for: calibration is work, and the rubric has to exist first, so this is a department decision more than an individual one. It has no reviews in our directory yet.
Pricing: "Free for teachers"; the page does not say what the free plan includes or limits. Schools and districts "can talk with our team," and the vendor's pricing address did not load: no public pricing. Checked November 16, 2026.
Safety: the privacy policy says "No AI provider we use has the right to train models on any data processed on EnlightenAI's behalf," and separately that the company may use de-identified information "to develop, research, and improve our Services" (privacy). It describes itself as a "School official" under FERPA, says it retains personal information until a teacher or school asks for deletion, and commits to deleting within 30 days of the request. The policy names no hosting location; our directory record lists United Kingdom hosting, a minimum age of 8 and a 9ine assessment. The vendor links a data processing agreement, which we did not read.
Olex.AI: thirty essays in one pass, handwriting included
Olex.AI is the best fit for an English department running frequent mock exams, because its SnapSnap feature "marks a whole class of 30 essays at the same time" and its OlexReader converts photographed handwriting to text (features). The vendor says it assesses English writing "from Key Stage 1 to Key Stage 5," and that at Key Stage 4 it "also marks whole GCSE exam papers in Science, Maths, Geography and History" (FAQs). Key stages are the UK's grade bands; Key Stage 4 is roughly grades 9 and 10.
Watch out for: it is "aligned with UK curriculum standards," so a US teacher is adapting rather than adopting. No free tier is stated and no price is published. Our directory holds two reviews, rated 4.00.
Pricing: "flexible pricing based on school size and use case (e.g. department, whole-school, or MAT deployment)," by personalized quote (a MAT is a multi-academy trust, a group of UK schools); a demo using your own students' work and a "bespoke trial" are offered (FAQs), checked November 16, 2026.
Safety: the vendor describes itself as "fully UK GDPR-compliant," says data "can be deleted at any time at your request," and makes the training statement quoted above (FAQs). Neither the FAQ nor the privacy policy names a hosting location; our directory record lists United Kingdom hosting and a minimum age of 4. UK GDPR is the right framework for UK schools. A US district is asking a different set of questions and should ask them directly.
Grade Direct: a published price and a credit meter
Grade Direct, formerly Graded Pro, is the best fit for a teacher who wants to know the cost before a sales call, because its site prints the figures: Teacher Free at $0 with "150 Credits Free" and "50 FREE Credits / Month - Max 50," and Teacher Pro at $25 for 6,000 credits with "No Expiration Date." It says it covers "every subject at every stage," and the teacher can "preview and edit feedback, add voice or written notes" before anything returns to students.
Watch out for: credits are an unfamiliar unit. The vendor's own table puts a basic task at about 3 credits, an in-depth task at 8 to 10 and grading a full exam paper at about 30, so count a month of your own assignments before buying. The page does not say whether $25 is a one-time purchase or a subscription, and its "less than $10 per month" line does not match its own 125-student example, as noted above. The old graded.pro address redirects to grade.direct, and the tool is no longer listed in our directory.
Pricing: Teacher Free $0; Teacher Pro $25 for 6,000 credits; School Account by custom quote with pooled credits (pricing), checked November 16, 2026.
Safety: the privacy policy, marked "Effective as of 9/5/26," says "We do not use Pupil Data to train AI models, and our terms with OpenAI confirm that submitted work is not retained by OpenAI for training" (privacy). It names its sub-processors: OpenAI, DigitalOcean for hosting (United States, with EU data center options), Google if you connect Google Classroom, Stripe and Mailgun. The policy does not say which region a given account uses, and it gives no minimum age. It also says that when a teacher acts for a school, "real Pupil Data should be routed through a School Account," which matters if you planned to use the free individual tier with named students.
Grade It AI: grading against a named exam board's mark scheme
Grade It AI is the best fit for a secondary English teacher grading to a specific specification, because it works to "AQA, Pearson, Cambridge, your assessment objectives, your bands" and takes work "handwritten or typed, photos or PDFs" (site). The vendor describes it as "an AI marking tool for English teachers and tutors."
Watch out for: it is scoped to English literature and language at secondary level, so it is the wrong tool for anything else. Gradings are metered per seat per month, so count one month's essays against the allowance before subscribing. Our directory record still lists it as contact-for-pricing, which is out of date.
Pricing: Starter $10.75 per seat a month, "billed yearly $129," for 30 gradings a month; Pro $20.75 per seat a month, "billed yearly $249," for 80. A top-up of 20 gradings costs $7.99. A founding offer of $99 for a year of Starter is open to the "first 50 founding seats." No free trial is stated; the homepage lets you grade one essay with no sign-up (pricing), checked November 16, 2026. The month-to-month prices are unverified, as noted above.
Safety: the privacy policy, last updated June 11, 2026, says "We do not sell personal data and do not use your content to train models of our own," and that uploads "are processed by the AI model providers we use for marking (currently OpenAI and Google)" (privacy). It does not say what those providers may do with the essays, so ask. Storage is on Amazon Web Services with no region named, where our directory record says European Union. The policy names no specific regulation and no minimum age; submissions are deleted "on account deletion or on request."
What should you ask a vendor before student essays go through it?
Copy these five questions into an email. Each one comes from a gap we found on the pages above.
- Is student writing used to train or improve any AI model, yours or your AI provider's? Please point to the sentence in your policy.
- In which country and region is our students' work stored, and does that differ by account type?
- Will you sign our district's data privacy agreement, and does it cover the free teacher plan or only a paid school plan?
- Which outside companies process the essays (for example OpenAI or Google), and what do your terms with them say about retention?
- How do we delete a class's work, how long does it take, and is anything de-identified kept afterward?
Our security assessments page shows which listed tools have an independent 9ine assessment on file, and AI tools a school can approve walks through the approval decision itself.
Is AI feedback as good as a teacher's?
Two 2026 studies are worth more to this decision than any vendor page.
On quality, a study of 979 authentic student responses across six STEM courses, with 472 students evaluating both versions, found that "the pedagogical quality of AI-generated feedback is, in practice, comparable to that of human-authored feedback," with only small effect sizes separating them. Human feedback carried more strategy guidance; AI feedback carried more metacognitive and motivational content. Neither was clean: roughly 13 percent of human feedback and 10 percent of AI feedback contained errors or ambiguities (Computers and Education: Artificial Intelligence, June 2026). The finding teachers should sit with is a different one from the same study: students' judgments "are largely influenced by their perceptions of the feedback provider's credibility rather than the actual quality of the feedback itself."
On risk, an experiment with 214 university teachers gave four groups the same medium-quality paper with different fabricated AI-detection reports. Where the report showed a high detection rate with red highlighting, mean quality ratings fell to 4.26 out of 10 against 6.83 for the same paper with a low detection rate and no highlighting; percentage scores fell from 73.30 to 51.61 (Frontiers in Psychology, July 7, 2026). The two middle conditions separate the two effects: the same paper rated 5.02 on a high detection rate with no highlighting, and 6.50 on a low rate with red highlighting, so the detection number does most of the damage and the red ink compounds it. The paper had not changed. Only the number next to it had.
That is the strongest practical argument on this page for a rule your department can adopt this term: keep detector output away from the person assigning the grade. None of the six tools above is an AI detector, which is part of why they belong in a feedback workflow and detectors do not.
One survey figure deserves the same plain treatment. In the Walton Family Foundation and Gallup study of 2,232 US public K-12 teachers, fielded March and April 2025, the share saying AI improved the quality of their work ranged from 57 percent to 74 percent by task, and grading and feedback sat at the bottom of that range, at 57 percent (Gallup, June 24, 2025). Teachers rate AI's help with feedback lower than its help with anything else they use it for. Weekly users in the same study estimated saving 5.9 hours a week, about six weeks a school year.
What does a faster feedback cycle look like over two weeks?
Here is how these tools fit a real unit, written as a plan rather than a promise. Thirty students, grade 9 English language arts, one argumentative essay of five paragraphs on a local issue.
Day 1 to 3, drafting. Students write in Google Docs. Brisk Boost is available for glow-and-grow feedback on demand, and students are told plainly what it is: a first reader, not a grader. No teacher time is spent here.
Day 4, a single checkpoint. You run rubric-criteria feedback across the class set from inside the documents. You read the feedback, not the essays, and you are looking for one thing: the three or four criteria the class is failing together. That is tomorrow's mini-lesson and it is worth more than thirty individual comments.
Day 5, teach the gap. Whole-class instruction on whatever the checkpoint surfaced, usually evidence-to-claim linkage at this level.
Day 6 to 8, revision. Students revise with the feedback tool still available. The revision, not the first draft, is what you will grade.
Day 9 to 10, grading. Final drafts go through the tool once for a proposed score against your rubric. You review every one before release, and you overrule it freely: the 10 percent error rate in the research above is roughly three essays in this class. Your comments go on the two or three pieces where a human reader changes what a student does next.
The unit spends the AI on the mechanical pass and spends you on the judgment. If your criteria are not written down yet, start with an AI rubric generator that applies your own standards, or see the grading tools that use your own rubric, then bring the rubric here. If the final drafts are on paper, the shortlist of tools for grading handwritten work covers what each can read.
What about Grammarly, Turnitin, MagicSchool and ChatGPT?
Grammarly is a writing assistant for the student more than a feedback tool for the teacher. Its Authorship feature is the part worth knowing: launched in October 2024 as a beta in Google Docs, it records whether text was "typed, pasted, or edited with AI tools." Grammarly was explicit that this is not detection, writing of detectors that "these algorithms are always going to trail the innovation driven by large language model (LLM) providers" (Grammarly, October 15, 2024). Its current availability was not checked for this update.
Turnitin is an integrity and similarity platform many districts already license. Its Feedback Studio page offers "drag-and-drop, voice, and text comments" that the instructor writes (Turnitin); the page we read on November 16, 2026 does not describe AI-drafted feedback, so do not buy it expecting that.
MagicSchool lists Class Writing Feedback among the key features of its free plan at $0, with Plus at $12.99 a month or $8.33 a month billed annually (pricing, checked November 16, 2026). If your school already runs MagicSchool, try its feedback tool before adding another subscription.
ChatGPT has a free plan and a Plus plan at $20 a month (pricing, checked November 16, 2026), and it will draft feedback from a pasted rubric. It keeps no class roster, no rubric library and no record of who has been told what, which is the administrative layer the six tools above exist to provide. Is ChatGPT enough for teachers? takes that question on its own.
Where should you start this week?
Start free and start where your students already write. For most teachers reading this that means Brisk Teaching this week, at no cost, on one class. If your department grades against a shared rubric, the calibration EnlightenAI is built around is the upgrade that pays. If you are grading a hundred handwritten mock papers in the UK system, Olex.AI is the one built for that job.
And keep the detector score out of the room where grades are decided. The research above found the same paper losing two and a half points out of ten because of a number printed beside it. Your judgment is the thing worth protecting, and it is the thing these tools are supposed to free up.
If what you need is grading at scale rather than feedback on writing, that is a different shortlist: see AI grading tools for teachers. English teachers looking more broadly should start with AI tools for English teachers.
Dan Fitzpatrick is a Forbes contributor, bestselling author (The AI Classroom, AI for School Leaders, The AI Educator), and founder of The AI Educator. He has trained over 150,000 educators across 100+ countries and advises schools, districts, and government bodies on AI strategy.
Change log. November 16, 2026: every price, free tier and data statement reread at source; table moved to the top with a verdict; training-data column, vendor questions and the October-to-November changes added; the claim that only Olex.AI states a no-training position corrected; the dead Graded Pro directory link replaced. October 18, 2026: retitled, opening shortened, and the claim that Grade Direct is the only tool with a published price corrected. October 5, 2026: prices first checked at source.


