New York's APPR is moving from the legacy 3012-d (HEDI) system to the new STEPS framework (Education Law 3012-e), built on the seven NYS Teaching Standards. Where AI helps with the observation write-up, where it doesn't, and why STEPS weighting and labels are set locally.

How can AI help New York evaluators write teacher evaluations under APPR and STEPS?

July 29, 202612 min read

How can AI help New York evaluators write teacher evaluations under APPR and STEPS?

New York's APPR is moving from the legacy 3012-d (HEDI) system to the new STEPS framework (Education Law 3012-e), built on the seven NYS Teaching Standards. Where AI helps with the observation write-up, where it doesn't, and why STEPS weighting and labels are set locally.

New York evaluates teachers through the Annual Professional Performance Review (APPR), and APPR is in the middle of its biggest change since 2010. Under Chapter 143 of the Laws of 2024, two systems now run side by side: the legacy Education Law §3012-d, with its single overall HEDI rating, and the new Education Law §3012-e — the Standards-based Educator Evaluation and Professional Support system, or STEPS — built on the seven New York State Teaching Standards. AI can meaningfully help with the observation write-up under either system: mapping evidence to the right standard and drafting standards-aligned feedback. What it can't do is decide which system your district uses, set your plan's weighting, or pick your rating labels — because under STEPS those are negotiated locally, not prescribed by the state.

What New York's evaluation landscape looks like right now

For years, APPR meant Education Law §3012-d: a state-prescribed structure, enacted in 2015, that combined classroom observation with other measures into a single overall HEDI rating — Highly Effective, Effective, Developing, or Ineffective. Districts negotiated some pieces locally, but the shape of the system, including the HEDI matrix, came from the state.

That changed on June 28, 2024, when Governor Hochul signed Chapter 143 of the Laws of 2024. The law amends §3012-d and establishes a new §3012-e, creating the Standards-based Educator Evaluation and Professional Support (STEPS) system. The Board of Regents adopted implementing regulations (Subpart 30-4) in March 2025. STEPS is a genuine philosophical shift: it removes the state-mandated test-score requirements of the old model, returns significant design authority to districts and unions, and reframes evaluation around professional growth aligned to the New York State Teaching Standards.

The transition is not a hard cutover. Districts and BOCES may keep their currently approved §3012-d plans — and submit material changes to them — through the 2031-32 school year, and must adopt a STEPS plan by June 30, 2032. In practice that means, for the next several years, some New York teachers will be evaluated under a 3012-d HEDI plan and others under a locally designed STEPS plan. Which one applies to you is a local decision, and it's the first thing any evaluator or vendor should confirm before assuming a framework.

The seven New York State Teaching Standards

STEPS is a standards-based system, and the standards it is based on are the seven New York State Teaching Standards, adopted by the Board of Regents in 2011. A STEPS plan must include multiple measures that together address all of the state's Teaching Standards, with at least one measure being classroom observation. The seven standards are:

Standard I — Knowledge of Students and Student Learning. Teachers acquire knowledge of each student and demonstrate knowledge of student development and learning to promote achievement for all students.

Standard II — Knowledge of Content and Instructional Planning. Teachers know the content they teach and plan instruction that ensures growth and achievement for all students.

Standard III — Instructional Practice. Teachers implement instruction that engages and challenges all students to meet or exceed the learning standards.

Standard IV — Learning Environment. Teachers work with all students to create a dynamic learning environment that supports achievement and growth.

Standard V — Assessment for Student Learning. Teachers use multiple measures to assess and document student growth, evaluate instructional effectiveness, and modify instruction.

Standard VI — Professional Responsibilities and Collaboration. Teachers demonstrate professional responsibility and engage relevant stakeholders to maximize student growth, development, and learning.

Standard VII — Professional Growth. Teachers set informed goals and engage in continuous, relevant professional growth.

These seven standards are the authoritative, statewide backbone of STEPS. What sits on top of them — the specific measures, how they're weighted, and what the rating labels are called — is where the local design happens.

The most important thing to understand about STEPS: it's locally designed

This is the feature of STEPS that trips up both evaluators and software, and it's worth stating plainly. Under Education Law §3012-e, a STEPS plan is locally designed and collectively bargained. The specific measures a district uses, how those measures are weighted, and the rating labels and descriptors are negotiated between the district and its union, then submitted to NYSED for approval. They are not handed down as a single statewide scale the way the old HEDI matrix was.

Practically, that means two neighboring New York districts can both run compliant STEPS plans that look quite different: different measures, different weighting, and potentially different rating language. Some plans may keep HEDI-style labels out of familiarity. Others may use the state framework's Level 1-4 numbering. Others may negotiate their own descriptors. All of those can be valid, because the statute deliberately pushes that design down to the local level.

For anyone writing evaluations — or building a tool to help — the implication is the same: the seven Teaching Standards are fixed and knowable, but the scale and weighting are local facts you have to confirm from your own district's approved plan. A tool can legitimately draft against the standards; it cannot legitimately invent your district's adopted scale.

After the observation: where the work piles up

Whichever system a district runs, the observation itself is familiar work for an experienced evaluator. The write-up that follows is where the hours go.

A single evaluation asks the evaluator to take fragmentary observation notes and turn them into evidence organized against the relevant standards, decide a rating for each, and draft feedback in language the teacher can act on — then do it again across a caseload, multiple times a year. Under 3012-d that means rendering evidence into the HEDI framework; under STEPS it means aligning evidence to the seven Teaching Standards and to whatever scale the district adopted. Either way, the observation takes a class period and the documentation takes considerably longer.

Where AI helps with the New York write-up

The strongest fit between current AI and APPR work is the same mapping-and-translation problem that shows up in every state framework, and it applies cleanly to New York's standards-based approach.

AI handles three parts of that reliably. First, it turns fragmentary notes into coherent, evidence-anchored prose in professional evaluation language — the difference between "kids explained their reasoning to each other" and a sentence tied to Standard III on instruction that engages students in constructing understanding. Second, it maps evidence to the relevant Teaching Standards, including evidence that supports more than one — a well-run discussion can touch Standard III (instructional practice) and Standard IV (learning environment) at once. Third, it maintains consistency across a full caseload, so the same quality of evidence lands in similar territory from teacher to teacher rather than drifting with the hour of the night.

For New York specifically, the value is in holding the write-up to the actual Teaching Standards rather than a generic "good teaching" gradient — and in doing it consistently enough that the documentation holds up across a faculty.

Where AI doesn't help — and what stays with the evaluator

The honest scope is shaped directly by how New York's systems are built.

AI cannot tell you which system or plan your district uses. That's a local decision — a 3012-d HEDI plan or a locally negotiated STEPS plan — and only your district can confirm it.

AI cannot set or calculate your APPR rating. The overall summative rating combines observation with your plan's other measures according to the weighting your district adopted. Under STEPS that weighting is locally bargained; the combination happens in your district's system, not in a drafting tool.

AI cannot invent your district's adopted scale. Because STEPS labels and descriptors are negotiated locally, a tool that assumes a fixed statewide scale will be wrong for some districts. The right approach is to draft against the fixed, knowable part — the seven Teaching Standards — and leave the labels adjustable to the plan you actually adopted.

AI cannot supply the local and professional context a seasoned New York evaluator brings — how a specific plan was negotiated, what a district's committee actually weighs, how a teacher's year unfolded. That judgment stays with the evaluator.

What AI does well, it does well. Pasting notes into a general chatbot works as a glorified search engine — but for standards-aligned documentation that maps to the seven Teaching Standards, holds a consistent scale, and survives review, the gap between "sounds polished" and "actually defensible" is wider than the polish suggests. General tools also have no built-in sense that New York now runs two systems, or that STEPS scales are local, so they tend to assume a statewide HEDI matrix that a given district may no longer use.

Common mistakes as districts move from 3012-d to STEPS

A few patterns are worth naming as districts transition.

The first is assuming the framework without checking. With two systems valid at once and STEPS plans varying by district, it's easy to default to the HEDI matrix everyone knows — even where a district has already adopted a STEPS plan with a different scale.

The second is treating STEPS as a fixed statewide rubric. It isn't. The seven Teaching Standards are fixed; the measures, weighting, and labels are local. Copying another district's STEPS scale, or a vendor's default, without confirming your own approved plan is a real risk.

The third is conflating the observation with the overall rating. The observation is one measure. Under both systems, the summative rating combines it with other measures under the plan's weighting — treating an observation rating as the final verdict skips that step.

The fourth is losing the growth purpose. STEPS was explicitly designed around professional growth and support rather than compliance. Documentation that reads as a checklist misses what the system is asking evaluators to do.

How EvalScribe handles New York's two systems

EvalScribe supports New York under both APPR systems. For §3012-d, it drafts observation ratings and comments against the HEDI scale your approved plan uses. For STEPS, it is built around the seven New York State Teaching Standards — the authoritative backbone every STEPS plan must address — and uses common HEDI-style labels (Highly Effective, Effective, Developing, Ineffective) as a starting convention for the ratings. An evaluator captures notes by typing, dictating, or photographing handwriting (Smart Scan OCR converts it to text), EvalScribe maps that evidence to the relevant standard, and drafts standards-aligned ratings and feedback. You review, edit, and finalize every rating and comment.

One honest scope note, because New York's design demands it. Under Education Law §3012-e, a STEPS plan's specific measures, their weighting, and its rating labels are set locally through collective bargaining and approved by NYSED — not prescribed by the state. Your district's adopted scale may use HEDI-style labels, the state framework's Level 1-4 numbering, or its own negotiated descriptors. EvalScribe drafts the standards-aligned observation write-up; confirm your district's adopted rating scale and weighting and adjust the labels to match your approved plan. EvalScribe drafts the observation piece — it does not set or calculate your plan's overall weighting or summative rating.

Beta testers report saving 30 to 60 minutes per evaluation versus writing the documentation by hand. Across a full caseload and multiple observation cycles, that adds up to dozens of hours back — hours that can go to the feedback conversation the ratings are meant to support. More detail on the New York workflow is available at evalscribe.com/newyork.

Frequently asked questions about APPR, STEPS, and AI

What is STEPS, and how is it different from 3012-d? STEPS — the Standards-based Educator Evaluation and Professional Support system — was established by Education Law §3012-e under Chapter 143 of the Laws of 2024. It replaces the state-prescribed structure of §3012-d with a locally designed, standards-based approach built on the seven NYS Teaching Standards, using multiple locally selected measures (at least one being observation). The older §3012-d system combined observation and other measures into a single overall HEDI rating under a more state-prescribed structure.

Is my 3012-d plan still valid? Yes. Districts and BOCES may keep their currently approved §3012-d plans — and submit material changes to them — through the 2031-32 school year, and must adopt a STEPS plan by June 30, 2032.

Who sets a STEPS plan's rating labels and weighting? Your district and its union, through collective bargaining, with the plan then approved by NYSED. The measures, their weighting, and the rating labels are set locally, not prescribed by the state — so confirm your own plan's adopted scale.

Does AI replace evaluator judgment? No. The observation, the professional judgment, and the rating decisions are yours. AI's legitimate role is translating your judgment into standards-aligned documentation.

Does EvalScribe calculate my APPR summative rating? No. EvalScribe drafts the standards-aligned observation write-up. The overall summative rating combines observation with your plan's other measures according to your district's adopted weighting, and that happens in your district's system.

Does EvalScribe support both systems? Yes. You choose STEPS or 3012-d when you set up an evaluation. For STEPS, EvalScribe is built on the seven Teaching Standards and uses HEDI-style labels as a starting convention that you adjust to your district's adopted scale.

If you're evaluating teachers in New York under APPR, see how EvalScribe handles both the STEPS and 3012-d workflows at evalscribe.com/newyork. Questions, or a school or district license? Reach the team at [email protected].

References

Related articles


Page maintained by Anthony D. Neely, Ph.D. — practicing K-12 educator with nearly 20 years in the classroom, 2025–2026 Walker County Distinguished Teacher of the Year, and co-founder of EvalScribe. Framework details verified against the New York State Teaching Standards, NYSED guidance, and Education Law §3012-d and §3012-e as amended by Chapter 143 of the Laws of 2024.

Last updated on July 28, 2026.

Anthony D. Neely, Ph.D.

Anthony D. Neely, Ph.D.

Anthony Neely is the Founder of EvalScribe, a veteran educator, an AI integration consultant for teaching & learning, researcher, & author.

Back to Blog