Six tools earn a place in an intervention teacher's week this fall: FrenalyticsEDU for progress-monitoring records, Diffit for practice text at a student's reading level, Conker for the weekly check, MagicSchool for the plan and the family letter, Kami for putting materials in front of students who need them read aloud, and Read&Write for the child who cannot yet decode the passage in front of them. Every one of them writes, levels, drafts or records. Not one of them decides who moves into Tier 2. That line is now written into state guidance, and it is the line this post is organized around.
The line your state just drew, and why it lands in September
Washington DC's Office of the State Superintendent of Education released a model AI policy for the 2026-27 school year in early September, sorting staff uses of AI into red, yellow and green (OSSE release). Red, meaning prohibited: student discipline decisions, teacher evaluation, physical surveillance, and IEP eligibility or Section 504 determinations. Yellow, meaning allowed with safeguards: drafting IEP language, reviewing and grading student work. Green, meaning permitted with human review: lesson planning, customizing student materials, developing tutoring plans, analyzing data sets, school communications.
Read that list as an intervention teacher, or as a SENCO running the graduated approach if you are in England, and the shape of your job appears in it. The eligibility call is red. The plan you write, the passage you level, the spreadsheet of fall screening scores you are staring at right now: green. The judgment stays with you. The typing does not have to.
The policy is a model, not a mandate, and it applies to DC. But it is the clearest published sorting of AI tasks in education this year, and district AI committees across the country are working from documents that look like it. If yours has not published one yet, this is a serviceable place to start the conversation.
How this list was built
The pool was the 28 tools in the directory's special education and accessibility category, plus every tool across the full 126-tool directory whose listing mentions intervention, progress monitoring or tiered support. I picked on four things: whether the tool does a job an intervention teacher actually has, whether it works without a student account where possible, what the vendor documents about student data, and whether it leaves the instructional decision with the teacher.
Nobody pays to be on this list and vendors cannot buy placement. Pricing and features were verified September 8, 2026.
Two well-known names were left out on purpose. Purpose-built MTSS platforms such as the ones your district may already license are not in the directory and are not covered here; they are procurement decisions, not tools a teacher adopts on a Tuesday. And I left out general-purpose chatbots as a headline pick, because the moment you paste a roster of screening scores with names attached into a consumer chatbot, you have made a student-records decision your district almost certainly has not approved.
What the directory shows about this corner of the market
Here is the finding that surprised me, and it is worth sitting with. Of the 126 tools in the directory, exactly one names MTSS anywhere in its listing. None name RTI, response to intervention, or universal screening. Two mention progress monitoring.
That is not a gap in our records. It is a gap in the market. The AI education boom has built lesson planners, quiz makers and feedback engines by the dozen, and has barely touched the documentation load carried by the people running Tier 2 and Tier 3.
The category has other things worth knowing. Of its 28 tools, 14 declare FERPA compliance and 12 declare COPPA, 4 state nothing at all about either, and 5 hold an independent 9ine security assessment. Eleven host data in the US, seven in the UK, and five do not say where. Those declarations are the vendor's own claim, not an audit; the 9ine column is the independent one.
The six, compared
| Tool | Best for | Price (checked Sept 8, 2026) | Free tier reality | Privacy (vendor-declared) | Independent assessment |
|---|---|---|---|---|---|
| FrenalyticsEDU | Progress-monitoring records | Free plan $0; paid from $1,499 | 10 student profiles | FERPA, COPPA; hosted US | No |
| Diffit | Practice text at level | $14.99/mo or $149.99/yr | 90-day access to what you make | Vendor states no student data collected | No |
| Conker | Weekly progress checks | No public pricing | Quiz building, limits not published | Not stated in our record | Yes (9ine) |
| MagicSchool | Plans and family letters | Free; Plus $8.33/mo billed annually | Limited generations, no saved history | Vendor states no training on student data | Yes (9ine) |
| Kami | Accessible delivery of materials | Free; Teacher $149/yr | 10+ of 40+ tools | FERPA, COPPA, GDPR; hosted US | Yes (9ine) |
| Read&Write | Decoding support during the block | Free for K-12 teachers; $184/yr individual | Full premium, teacher account only | FERPA and COPPA badges on vendor site | Yes (9ine) |
Best for progress-monitoring records: FrenalyticsEDU
FrenalyticsEDU is the best fit for keeping intervention data in one place because it is the only tool in the directory built for that job. Its listing describes automated progress monitoring, IEP goal tracking and skill-based assessments, and it names MTSS, 504 and English learner caseloads alongside special education.
What sets it apart from a shared spreadsheet is that the data collection and the goal are attached to each other. A student works through a personalized activity, the platform records what happened, and the record maps to the goal rather than sitting in a separate tab you update on Friday afternoon.
Watch out for: this tool holds real student records. Its directory card shows it collects names, dates of birth and performance data, hosts in the US, and does not share with third parties. That is a district conversation and a signed agreement, not a free trial you start on your own. The pricing page also displays its paid plans with a monthly label while stating that billing is annual after a two-week trial, so confirm the period before you take a number to your principal.
Best for: a caseload you already document formally. Not for: a teacher who wants to try something this week without involving the district.
Pricing: a free plan at $0 covering 10 student profiles with no trial and no card required; EDU Essentials listed at $1,499 for 15 profiles and EDU Extras at $1,699 for 20, with the billing period shown ambiguously and therefore unverified (pricing, checked September 8, 2026).
Safety: the vendor declares FERPA and COPPA compliance and US hosting, and states no minimum age. Because it stores identifiable student performance data, your district will need a written agreement with the vendor before you enter a single name. The Department of Education's Privacy Technical Assistance Center publishes a written agreement checklist covering what such an agreement has to specify, including limits on use and destruction of data when the purpose ends.
Best for practice text at level: Diffit
Diffit is the best fit for producing Tier 2 practice text because it takes a passage you already teach and returns it at the reading level a student can work in, with questions and vocabulary attached. Paste a URL, a PDF or a chunk of text; choose the level; get a resource in about a minute.
Its school plan explicitly includes MTSS tier options and standards alignment, which is a rare thing to see named on a pricing page. That naming is also the catch: tier options sit behind the school plan, not the free one.
Watch out for: the free plan only lets you reach content you made in the last 90 days (pricing). If you build a leveled set for a six-week cycle and come back to it next semester, it is gone unless you exported it.
Best for: reading and content-area intervention where the text is the barrier. Not for: math fact fluency or anything that is not passage-based.
Pricing: free Basic plan; individual teacher $14.99 per month or $149.99 per year (subscription page); school pricing is a flat annual rate tiered by enrollment with no public figure (pricing, checked September 8, 2026).
Safety: our directory record is blank on FERPA, COPPA and hosting. The vendor's own privacy page states that it collects no student data and describes the service as FERPA and COPPA compliant, and it collects teacher names, emails and IP addresses only. Students never log in, which is what makes it usable while your district decides about everything else. The blank in our record needs fixing, and Dan has the note.
Best for weekly progress checks: Conker
Conker is the best fit for a quick progress check because it generates a standards-aligned quiz with read-aloud and an adjustable reading level, which is exactly what a curriculum-based measure needs and what most quiz generators leave out.
The reading-level control matters more here than anywhere else in this list. A progress check written two grades above a student's decoding level measures their decoding, not the skill you taught.
Watch out for: no public pricing was retrievable during this run, and our record carries nothing on FERPA, COPPA or data hosting. It does hold an independent 9ine assessment, which is more than most of the category can say. Treat the compliance question as open until your district checks it.
Best for: a short, repeatable check every week or two. Not for: the formal, validated progress-monitoring measure your district requires for tier decisions. This supplements that; it does not replace it.
Pricing: freemium. No public pricing was confirmed in this run (pricing page, checked September 8, 2026).
Safety: independently assessed by 9ine. FERPA, COPPA, GDPR and hosting are all not stated in our record and were not published in a form we could verify this run, so this one is unverified. Students take the quizzes, which means it is a district question rather than a teacher one.
Best for plans and family letters: MagicSchool
MagicSchool is the best fit for the writing around an intervention because it has the widest set of teacher tools in the directory, carries a 4.74 rating from 19 reviews on its directory page, and drafts the two documents that eat an intervention teacher's Friday: the plan and the letter home.
Give it the skill, the baseline, the goal and the schedule, and it returns a draft you rewrite. That word, rewrite, is doing all the work in this section.
Watch out for: in August 2025 Common Sense Media assessed four AI teacher assistants including MagicSchool and rated the category moderate risk. It found that behavior plan generators "can demonstrate problematic bias based on perceived student backgrounds," and that identical prompts differing only in whether a student's name read as Black or white produced more punitive recommendations for the Black-coded names (Common Sense Media). The report judged AI-generated IEP content inadequate for high-stakes use. Behavior plans and intervention plans are the same species of document. Take the structure from the draft. Supply the child yourself. The same caution runs through the IEP writing post, and it applies with more force here, because an intervention plan gets less scrutiny than an IEP on its way into a file.
Best for: the first draft of a plan or a family communication. Not for: anything that goes into a file without you rewriting it line by line.
Pricing: free plan with limited generations and no saved output history; Plus at $8.33 per month billed annually or $12.99 monthly; Enterprise by quote (pricing, checked September 8, 2026).
Safety: the privacy policy states that MagicSchool does not use student data to train or fine-tune models and that its third-party AI providers are contractually barred from doing so either. It holds a 9ine assessment. Our directory record is blank on FERPA, COPPA and hosting, and the Enterprise tier is where custom data privacy agreements live, so a district deployment is a different conversation from a teacher account.
Best for accessible delivery of materials: Kami
Kami is the best fit for putting intervention materials in front of students because it turns a worksheet or PDF into something a student can annotate, hear read aloud and work in, and its directory card carries the cleanest privacy record of the six: vendor-declared FERPA, COPPA and GDPR, US hosting, no minimum age, no student data collected, deletion on request, and a 9ine assessment.
Where Diffit makes the text, Kami delivers it. A leveled passage exported as a PDF becomes a workspace with text-to-speech and annotation, which is often the difference between a student in a small group participating and a student watching.
Watch out for: the free Basic plan gives you 10 or so of the 40-plus tools, and the read-aloud and assessment features people actually want sit above it. Check what is in Basic before you build a routine on something you will lose (free plan detail).
Best for: getting materials to students with reading or motor barriers. Not for: creating the content in the first place.
Pricing: free Basic plan; Teacher $149 per year; school and district by quote (pricing, checked September 8, 2026).
Safety: vendor-declared FERPA, COPPA and GDPR, hosted in the US, with our record showing no student data collected and deletion available on request, plus an independent 9ine assessment. Still confirm your district's approved-tools list, because student accounts are involved.
Best for decoding support during the block: Read&Write
Read&Write is the best fit for the student who cannot yet read the intervention material because it does one thing well and has done it since 1996: text-to-speech with dual-color highlighting, word prediction, vocabulary support and translation, across the browser and the desktop.
Texthelp, now Everway, gives K-12 teachers a free premium subscription (free for teachers). For a teacher deciding whether accessible text changes what a student can do in a twenty-minute block, that removes the cost of finding out.
Watch out for: the free offer covers your teacher account, not your students. Student licenses are a purchase, and the district and group pricing on the pricing page tells you this is bought at scale, not per child.
Best for: decoding and vocabulary barriers that block access to grade-level content. Not for: teaching decoding itself. Reading it aloud and teaching a student to read it are different jobs, and the tools that go further than reading aloud are covered in the post on AI tools for students with dyslexia.
Pricing: free premium subscription for K-12 teachers; individual license $184 per year; K-12 Group listed at $18.30 per seat with a 150-seat minimum and K-12 District at $2.75 per seat (pricing, checked September 8, 2026).
Safety: our directory record is blank on FERPA, COPPA and hosting. The vendor's privacy page displays FERPA, COPPA, GDPR, ISO 27001 and SOPPA compliance badges, and the underlying policy documents were not retrievable in this run, so treat the detail as unverified and ask for it in writing. It holds a 9ine assessment.
The names AI assistants will hand you, and where they actually fit
Ask ChatGPT or Perplexity for intervention tools and four names come back. Here is the honest one-liner on each.
Khanmigo is a tutor, not an intervention record. It is useful as the independent practice half of a rotation and it holds a 9ine assessment and a 4.40 directory rating from 5 reviews, but it will not document a thing for you (directory page).
ChatGPT is the best drafting partner on this list for the wording of a plan or a parent email, provided you strip every name and identifier first. Give it "a fourth grader reading at a second-grade level," never a roster.
NotebookLM is the right tool for the wrong-looking job: load your district's MTSS handbook and your state's guidance into it and ask what the fidelity requirements are. It answers from your documents, which is exactly what you want when the question is procedural.
Quizlet builds recall practice quickly and students already use it. It is not a progress-monitoring instrument, and treating a set score as data is how a Tier 2 decision gets made on the wrong evidence.
How to decide what AI touches in your intervention block
Four questions, in this order. They come from a rule worth stating plainly: outsource the doing, not the thinking. In intervention, the thinking is who needs what and whether it is working. Everything else is doing.
1. Does this task decide something about a child? Tiering, exiting, referring, eligibility. If yes, no tool touches it. This is the red list, and it is the shortest list, and it is not negotiable.
2. Does it need a name attached? If yes, the tool needs a district agreement first. If no, you can use it today. Most of what an intervention teacher does with AI, leveling a passage, drafting a plan template, writing a family letter about a generic six-week cycle, needs no name at all. Strip identifiers and a compliance problem becomes a Tuesday. The tools that work this way are gathered in the post on AI tools that need no student account.
3. Would I recognize a bad output? You will spot a wrong reading level in ten seconds. You may not spot a plan that quietly recommends a harsher response because of a child's name, which is what Common Sense Media found. The rule: use AI where your expertise catches the error, and slow down where it does not.
4. Does it replace an instrument my district relies on? A Conker quiz is a check. Your district's validated measure is the evidence. The IES practice guide on response to intervention in the primary grades recommends monitoring Tier 2 students at least once a month, roughly eight times a year, and regrouping about every six weeks (Gersten and colleagues, 2009). That cadence is met with your instrument. Everything here supplements it.
What I would set up in the first two weeks
Take one group. Say six fourth graders, twenty minutes, four days a week, working on inferential comprehension.
Week one: build the six-week text set in Diffit from passages you already teach, at two levels, and export every one as a PDF the same day so the 90-day window never bites you. Put those PDFs into Kami so the students who need audio have it, and hand the routine to whoever runs the group with you, which is often a paraprofessional. Draft the group plan in MagicSchool, then rewrite it with these six children in mind, because the draft does not know them.
Week two: make one Conker check on the week's skill at the group's reading level. Run it Friday. Record it beside your district's measure, not instead of it.
Then measure one thing at the six-week mark: did the number of students meeting the exit criterion go up, and did your Friday afternoon get shorter? If neither moved, the tools are not the problem and swapping them will not help.
The bottom line
For intervention teachers this fall: Diffit for the text, Kami to deliver it, Conker for the weekly check, MagicSchool for the paperwork, Read&Write for the students who cannot yet decode, and FrenalyticsEDU if your district will sign for a real progress-monitoring record. Compare the rest in the special education and accessibility category or across the full directory.
The market has built a hundred lesson planners and one tool that says the word MTSS. Until that changes, the job is to use good general tools carefully and keep the decision where it belongs. A screening score tells you something happened. It does not tell you what to do about a nine-year-old. That is still your work, and it is the part worth protecting.
Dan Fitzpatrick is a Forbes contributor, a three-time bestselling author and the founder of The AI Educator, and has trained more than 150,000 educators across 30 countries. More from Dan.
