AI & documentation

Can AI help write IEP progress reports?

Allie Wilson, Founder, IEP Assure · September 30, 2026 · 3 min read

Quick answer

Yes, if it drafts from evidence your teachers already recorded and an educator reviews and finalizes every word. AI should never decide eligibility, services, tiers or whether a service happened. Used that way, it gives teachers time back without taking judgment away from them.

Where AI helps

Progress reports are repetitive to write and easy to write poorly. A good draft pulls from the data points, session notes and work samples already logged for each goal, then organizes them into clear language a family can follow.

That draft is a starting point. The teacher knows the student; the draft simply saves them from starting at a blank page every six weeks.

Where AI does not belong

Deciding eligibility, placement or services.
Setting or changing a student's tier or level of service.
Inventing progress that is not in the record.
Replacing the ARD or IEP committee's judgment.

If a tool cannot show you which evidence a sentence came from, it should not be writing that sentence.

What a good draft looks like

A useful progress report draft does three things. It states where the student is on the goal using the same measure the goal uses. It compares that to the baseline and the last reporting period. And it describes what the team is doing next, in plain language a family can act on.

A weak draft does the opposite. It talks about effort and attitude instead of performance, uses general phrases that could describe any student, or reports numbers that do not appear anywhere in the data. Those are the drafts teachers end up rewriting from scratch, and they are the reason many teams are skeptical of AI in the first place.

The difference is almost always the input. When a draft is built from dated data points, session notes and work samples tied to each goal, the language has something real to describe.

Guardrails to require

1Drafts use only information your staff entered. No outside data about the student.

2Every draft is reviewed, edited and approved by an educator before it is shared.

3Student data is not used to train AI models.

4The review step is visible, so leaders can see what was approved and by whom.

5Your data agreement names the AI provider and how requests are handled.

Questions for your vendor

1Which AI provider do you use, and does that provider store our requests?

2Can a teacher see which logged evidence each part of the draft came from?

3What happens if there is not enough data to write a draft?

4Where is the approval step recorded, and who can see it?

5Is the no-training commitment written into our data agreement, not only the website?

Clear answers to these questions matter more than any feature list. If a vendor is vague on data handling, treat that as the answer.

What leaders should watch

The quality of an AI draft is only as good as the documentation underneath it. If service logs are thin, drafts will be thin. That makes daily documentation the real lever, and AI a way to get more value from it.

A practical rollout

Pilot AI drafting with one grading period and a small group of teachers who already log consistently. Compare the drafts to what those teachers would have written, and ask them where the draft helped and where it missed. Expand once the review step feels natural and the documentation underneath is strong. That order protects students, keeps educators in charge and builds trust with families.

See what this looks like in your district.

A 30-minute walkthrough built around your campuses and your questions.

Book a 30-minute walkthrough