Using AI to Transform IEP Writing: A Teacher’s Journey From…

Dec 22/Jan 23 Closing The Gap Solutions - Collaborating With Augmentative and Alternative Communication (AAC) Users Gains A New Perspective To Best Support Clients By Lydia Dawley

leadership, policy & implementation

Summary: Special education teacher Tim Ranis shares how artificial intelligence transformed his IEP writing process and restored his passion for teaching. After nearly three decades in the field, crushing paperwork drove him toward early retirement until he developed KAIIT—the Kairo Artificial Intelligence IEP Tool. This article details the replicable six-step process demonstrated at Closing the Gap 2025, showing educators how to ethically reduce IEP writing time significantly while maintaining profes- sional oversight, legal compliance, and student-centered quality. Using AI to Transform IEP Writing: A Teacher’s Journey From Burnout to Breakthrough

INTRODUCTION: WHEN PAPERWORK ECLIPSES PURPOSE No special education teacher enters the profession dream- ing about paperwork. We come for the students—their poten- tial, their growth, the privilege of watching breakthroughs un- fold, the “Ah-ha” moments. Yet somewhere along the journey, the administrative load expands until it threatens to consume the very reason we chose this work. Last spring, after 27 years teaching special education, I reached an all too familiar stress point. IEP fatigue had become so profound that I was counting down months, then weeks, even days until retirement. The students weren’t the problem. The classroom wasn’t the problem. The suffocating weight of paperwork that followed me home every night and consumed every weekend—that was the problem. Then artificial intelligence entered my professional life. Not as a gimmick or shortcut, but as a genuine lifeline. The tool I developed—KAIIT (Kairo Artificial Intelligence IEP Tool)—has fundamentally reshaped how I write IEPs. It has restored my ability to focus on students and strengthened/ improved, rather than replaced, my professional judgment.

What follows is the story behind KAIIT, insights from the live demonstration I conducted at the Closing the Gap conference, and clear, replicable steps any educator can use to lighten their workload without compromising quality, accuracy, or compli- ance. FROM DESPERATION TO INNOVATION: THE EASTER EXPERIMENT KAIIT’s origin story lacks elegance. It wasn’t planned, fund- ed, or officially sanctioned. I made mistakes, many mistakes. It was born from desperation during Last year’s IEP season when I found myself buried under seemingly never-ending mountains of paperwork from student meetings. Over Easter break, I did something unusual: I spent nearly 40 hours experimenting with AI prompts, testing whether this technology could genuinely assist with IEP writing. Since I pri- marily support math as a Special Education teacher, I was ac- customed to logic and patterns, so I approached AI not with blind faith but with skepticism and systematic testing. What I discovered surprised me. AI could compile, sum- marize, organize, and synthesize information with remarkable

TIM RANIS is a Wisconsin special education teacher, retired Army National Guard Staff Sergeant, and creator of KAIIT (Kairo Artificial Intelligence IEP Tool)— the AI-driven workflow transforming IEP writing. After nearly three decades in the classroom, he developed KAIIT during the 2024 school year and presented this work at the Closing the Gap Assistive Technology Conference. Tim remains committed to translating cutting- edge innovations in Artificial Intelligence into practical, ethical support for special educators everywhere, as well as training educators to embrace and utilize AI in the classroom. He is currently working on pure research relating to the Distribution of Artificial Intelligence Identities across platforms.

BACK TO CONTENTS

YouTube Video - This 5 minute demonstration shows how to use the KAIIT 3.8 master prompt to streamline the IEP writing process. It will turn hours of work into minutes. https://www.youtube.com/watch?v=Y2lm9bhKtyU

clarity—if given clear instructions and high-quality input data. The technology wasn’t magic; it was a pattern-matching tool that excelled at exactly the kind of repetitive, data-intensive work that buries special educators - work that was burying me. Through extensive trial and error, I developed a structured method for guiding AI to behave like an experienced special education teacher who understands both legal requirements and classroom realities. I named this protocol KAIIT. KAIIT isn’t proprietary software or a commercial application. It’s a replicable pattern—a systematic approach that any edu- cator can implement using whatever AI platform their district approves. The genius isn’t in the technology itself; it’s in the process. THE CLOSING THE GAP DEMONSTRATION: WHEN THEORY MEETS REALITY Presenting KAIIT at Closing the Gap carried significant un- certainty. I had prepared materials, loaded documents, and hoped the technology would cooperate. What I didn’t antici- pate was the interactive reaction in that conference room. The moment AI produced four polished IEP paragraphs in sec- onds, educators physically leaned forward. I heard gasps, the

scraping of chairs being pulled closer. One participant later described watching the AI compile multiple data sources as “seeing the future.” Another admitted, “I didn’t believe this was possible until I saw it happen.” Perhaps most telling: when the session ran 15 minutes over schedule, not a single person left. (Sorry 11:30 session.) That retention wasn’t about my presentation skills—it reflected the desperate need in our field for sustainable solutions to over- whelming workloads. The demonstration proved something crucial: this isn’t the- oretical. It’s practical, ethical, and immediately applicable. THE KAIIT PROTOCOL: SIX STEPS TO RECLAIMING YOUR TIME STEP 1: GATHER COMPREHENSIVE STUDENT DATA AI functions only as well as the information you provide. When data is absent, AI may fill gaps with hallucinations—fab- ricated content that sounds plausible but lacks factual basis. Therefore, data collection is foundational. Before engaging with AI, I compile as many of these docu- ments as available and accurate:

BACK TO CONTENTS

STEP 4: REQUEST INDIVIDUAL IEP SECTIONS SEQUENTIALLY This procedural approach makes KAIIT both reliable and safe. You never ask AI to “write a complete IEP” because you will lose specificity, AI will skimp on the individual sections and increase the probability of hallucinations. Instead, you prompt for one section at a time: (My example student was Steve) • “Create the present level of academic achievement and functional performance for Steve.” • “Summarize the parent concerns expressed at the meet- ing.” • “Describe how Steve’s disability affects involvement in general education curriculum.” • “Please rewrite this output in one concise paragraph.” • “Expand this section into 10 paragraphs explaining greater detail.” • “Create a SMART goal based on the meeting discussion.” • “Place current grades in a table format.” • “List accommodations that support Steve’s learning needs.” The AI retrieves only information matching the specific pat- tern and question. As I emphasized during the demonstration: “I’m not asking it to do anything crazy or unpredictable. I’m ask- ing for exactly what belongs in that specific section of the IEP.” This approach works regardless of your state’s IEP format. You simply ask the AI to respond to whatever question appears in your particular documentation system. Whether you use SEEDS in Wisconsin, SIRAS in another state, a fill-in Word document or a different platform entirely, the principle remains constant: ask specific questions, receive targeted responses. STEP 5: EXERCISE PROFESSIONAL JUDGMENT THROUGH REVIEW AND EDITING AI handles approximately 90% of the compilation work (This is my anecdotal data, I need a larger sample size to quantify it), but educators, you, remain responsible for the critical final 10%. I cannot overstate this point’s importance. Throughout the demonstration, I repeatedly emphasized: “You are still The Professional. You are the one who makes the final decision. AI is doing the heavy lifting for you, but in the end, you’re the one putting this IEP forward.” Your name is on

• Draft IEP with last year’s narrative sections removed • Teacher input forms detailing current classroom perfor- mance and current grades • Student-led IEP PowerPoint or preparation notes • Meeting notes or transcripts • Assessment results and progress monitoring data • Standardized testing scores • Transition surveys or Postsecondary Transition Plans • Behavioral data, when relevant • Social-emotional learning profiles • Assistive technology interventions and rationale During the conference demonstration, I showed participants each document type, explaining: “This is what I prepare for an IEP meeting. Whatever data I have on the student, I drop it in.” The beauty of modern AI platforms is their simplicity. You liter- ally drag PDF files into the chat window. No special formatting required. No technical expertise needed. This accessibility often elicits the first gasp from educators who expected complexity. STEP 2: DEPLOY THE KAIIT MASTER PROMPT This instruction set forms the system’s heart. It transforms generic AI into a knowledgeable assistant that understands special education’s unique demands. Key elements from the master prompt include: “Assume the role of an experienced special education teach- er who writes IEPs daily and understands both legal and practi- cal classroom requirements. Treat this prompt as your source of truth. Create detailed, professional, legally grounded IEP narra- tive sections using a formal tone. Base all narratives on observ- able facts from the provided documents. Balance detail with clarity—be concise but never vague. If insufficient data exists to answer a question, indicate ‘data not provided’ rather than fabricating information. Provide output in paragraph form un- less the user requests tables or alternative formatting.” These instructions prevent the AI from drifting off-task, maintain appropriate professional tone, and keep output aligned with best practices. Teachers don’t need technical ex- pertise—the prompt handles the heavy lifting. STEP 3: ESTABLISH CONTEXT THROUGH INITIAL PROMPT Once documents are loaded, you issue one foundational in- struction that frames everything that follows: “We are going to create an IEP for a 10th grade student in Wisconsin . Please review the data” This single sentence tells the AI the grade level, state re- quirements, and purpose. The AI now understands context and can draw from appropriate legal frameworks and developmen- tal expectations.

this submission, this legal document. When reviewing AI output, you must:

• Verify accuracy against source documents • Remove irrelevant or incorrect information • Add missing details manually or through additional prompts • Correct any hallucinations or fabrications • Ensure descriptions match the student you know

BACK TO CONTENTS

Tim Ranis presenting at the Closing the Gap Conference in Minneapolis in October 2025.

STEP 6: MAINTAIN PRIVACY, SAFETY, AND COMPLIANCE FERPA obligations don’t disappear because you’re using new technology. Data privacy remains paramount. During the conference, multiple participants asked about security. I addressed concerns directly by explaining my mul- tilayered approach: First, I comply with district requirements. My school district ultimately required me to use Google Gemini rather than Chat- GPT because of enhanced data contracts with Google’s educa- tion division. These contracts include “hold harmless” clauses that provide additional legal protection for the district. Second, I adjust AI platform settings to prohibit training on my data. In ChatGPT, this means navigating to Settings > Data Controls > “Improve the model for everyone” and ensuring it’s turned OFF. This prevents your student information from being used to train future AI models. Third, I anonymize data as much as possible. In demonstra- tions, I use fictional names like “Captain America” or “Steve” rather than actual student names. Even in private IEP work, I replace names with only first names or initials when feasible.

• Adjust tone or phrasing that doesn’t sound authentic • Confirm legal compliance and required elements

During my live demo, the AI occasionally pulled information that didn’t quite fit. For example, it suggested Steve was inter- ested in career exploration, but meeting notes didn’t support this claim. I immediately deleted that phrase, demonstrating real-time editing judgment. This editing process preserves several essential elements: • Legal accuracy and compliance • Student individuality and authentic representation • Ethical responsibility to families • Professional ownership of the final document

The AI won’t be the professional submitting the IEP for ap- proval. You will. That’s why this step is non-negotiable.

BACK TO CONTENTS

Fourth, I communicate transparently with parents. When asking for permission to record meetings for transcription pur- poses, I explicitly ask, explain the AI use, and commit to delet- ing recordings by day’s end. Every parent I’ve approached has consented, but if someone someday declines, I’d immediately revert to traditional note-taking. Every district’s policies differ. Some allow consumer-grade AI platforms with appropriate settings. Others require enter- prise education versions with locked-down security. Know your district’s requirements and follow them without exception. WHY THE HUMAN-AI PARTNERSHIP WORKS AI HANDLES VOLUME; TEACHERS HANDLE TRUTH The paperwork burden in special education doesn’t stem from laziness or inefficiency. It stems from sheer volume. AI ex- cels at managing: • Repetitive summarization across multiple documents • Organizing multi-page assessment reports • Reformatting data into tables or narratives • Crafting grammatically clean paragraphs • Maintaining consistent tone across sections

• Observing student strengths and challenges • Building relationships with families • Collaborating with general education colleagues

• Delivering targeted instruction • Providing timely interventions

• Analyzing data for program adjustments • Maintaining healthy work-life balance

KAIIT doesn’t improve the teaching experience because AI possesses magic powers. It improves teaching experience be- cause teachers finally have time and mental energy to think deeply about the compiled information of the IEP and apply it thoughtfully to students they serve. AI PROVIDES CONSISTENCY THAT REDUCES ERRORS Humans get tired. We rush. We forget the required elements. We fail to include all the details from all the documents. We accidentally omit components or mix up formatting between students.

AI doesn’t experience fatigue. It consistently remembers:

• Required formatting elements • Appropriate professional tone • State-specific compliance components • Transition planning requirements • Standardized test reporting language

However, AI fundamentally cannot:

• Understand a student’s personality and learning style • Interpret nuanced parent concerns expressed during meetings • Determine individuals developmentally appropriate ma- terial • Comprehend school culture and community context • Make educational placement decisions • Ensure true legal compliance beyond surface-level for- matting You remain The Professional. You hold both the authority and the responsibility. This creates, “the partnership”: AI com- piles the data; teachers interpret and apply that data with wis- dom, experience, and care. STUDENTS BENEFIT WHEN TEACHERS RECLAIM TIME At the conference, I shared something I hadn’t publicly ac- knowledged before: “This literally changed my life. It got rid of the paperwork load [stress]. I can now work with kids instead of being anxious about paperwork.” That statement captures this work’s true heart. That’s why I give KAIIT away for free. When educators spend less time typing nar- rative paragraphs and formatting tables, they can invest more energy in:

This consistency prevents common omissions and saves teachers from rewriting nearly identical paragraphs across mul- tiple IEPs. More importantly, it creates a reliable baseline that teachers can then customize with personal knowledge and professional insight. STUDENT-CENTERED LANGUAGE AND PROFESSIONAL QUALITY The KAIIT master prompt specifically instructs AI regarding person-first language: “Use person-first language unless iden- tity-first terminology is explicitly appropriate for the individual or community.” Sometimes we struggle with how to phrase this. AI does not. This produces respectful, strength-based descriptions such as: • “A student who benefits from structured routines and visual schedules” • “A learner who uses assistive technology to access grade- level curriculum” • “An individual with autism who identifies as Autistic and advocates for neurodiversity awareness”

BACK TO CONTENTS

The prompt also requires observable, evidence-based state- ments rather than vague generalizations:

told participants, “IEP writing has become achievable [at that level].”Not because the technology dazzled with special effects, but because it proved accessible. With an AI platform, quality input data, and the KAIIT framework, hours of IEP writing can truly become minutes—and significantly fewer hours over the course of the year - while still producing quality. One educator whispered to her colleague, “This would save me hours every week.” Another approached afterward admit- ting,“I didn’t know we were allowed to do this.”You are allowed. You simply need to proceed responsibly, transparently, and with full professional oversight. FIVE RULES FOR SAFE, ETHICAL AI IEP WRITING 1. USE REAL DATA ONLY—NEVER ALLOW FABRICATION If AI invents information not present in your source documents, delete it immediately. Accuracy always trumps speed. 2. TREAT AI AS COMPILER, NOT DECISION-MAKER Your professional judgment remains final on every element: goals, placement, services, accommodations, and narrative de- scriptions.

• Instead of: “Steve struggles in class” • AI produces: “Steve’s current chemistry grade of F re- flects challenges with abstract concept application, as noted by Mr. Smith in the teacher input form” This grounding in specific, documented information cre- ates IEPs that withstand scrutiny and serve students more ef- fectively. THE CONFERENCE ROOM’S MOST POWERFUL MOMENT Midway through the demonstration, I paused to address the palpable anxiety in the room: “Don’t be scared of artificial in- telligence. Don’t worry about it. I’m going to help you use this tool.” What followed still resonates with me. When AI compiled multiple data sources into clean IEP paragraphs in mere sec- onds, then produced several complete sections within min- utes, the entire room shifted. The energy changed from skepti- cal curiosity to seriously engaged possibility. Compliant, high quality IEP writing has always been what we strive for, but time is our nemesis. With AI assistance, as I

Reference Card, a “Cheat Sheet” for following the KAIIT process of writing IEP’s with AI.

BACK TO CONTENTS

BEYOND IEP WRITING: EXPANDING KAIIT APPLICATIONS

3. BREAK IEPS INTO SECTIONS—NEVER REQUEST COMPLETE DOCUMENTS Asking AI to generate an entire IEP at once increases error risk and reduces your control. Sequential section requests maintain quality and oversight. 4. FOLLOW DISTRICT PRIVACY AND SECURITY REQUIREMENTS When uncertain about policies, err toward caution. Anonymize data, adjust platform settings, and communicate with adminis- trators, IT and Student Services about your process. 5. READ EVERY LINE BEFORE SUBMISSION AI accelerates writing but doesn’t reduce your responsibility. Review everything with the same care you’d apply to manually written content. ADDRESSING COMMON CONCERNS “WON’T THIS MAKE IEPS GENERIC?” The opposite often occurs. Because teachers have more time and mental energy, they can add personalized details, specific examples, and nuanced observations that make IEPs more in- dividualized, not less. The bulk of the individual data has been compiled by the AI. You are the cherry on top. “WHAT ABOUT STUDENTS WITH COMPLEX NEEDS?” KAIIT works equally well for students with intensive support requirements. The key is providing comprehensive data: behav- ioral plans, medical information, communication assessments, sensory profiles, and detailed progress monitoring. More data produces better, more tailored output. “DOES THIS VIOLATE THE SPECIAL EDUCATION SPIRIT?” KAIIT enhances special education’s core mission by freeing professionals to focus on students rather than paperwork. The spirit of special education lies in individualized support and family partnership - neither of which requires teachers to man- ually type nearly repetitive paragraphs. “WHAT IF MY DISTRICT PROHIBITS AI?” Start conversations with administrators. Share this article. Propose pilot programs with clear privacy protocols. Many dis- tricts have legitimate concerns about consumer AI platforms but may approve education-specific versions with appropriate contracts. Advocate for tools that support retention and reduce burnout in our profession. Keep your eyes open as AI changes continually. What may be considered insecure now, may be ac- ceptable tomorrow.

During the conference, participants asked whether KAIIT could assist with other tasks. The answer is yes, though I’m still developing some applications: IEP Snapshots: I load completed IEPs at the school year’s start and ask AI to create one-page summaries for general educa- tion teachers. This formerly time-consuming task now takes minutes. Even better: as one participant brilliantly suggested, students themselves could create their own snapshots as a self- advocacy exercise with AI. Progress Monitoring Forms: AI can generate customized data collection tools based on IEP goals, creating consistent tracking systems across caseloads. Parent Communication: AI helps draft meeting summaries, progress letters, and follow-up emails in professional language while maintaining personal warmth. The underlying principle remains constant: provide quality data, ask specific questions, and edit output with professional judgment. THE FUTURE OF AI IN SPECIAL EDUCATION I’m not naive about potential risks. AI can perpetuate biases presented in training data. It can produce plausible-sounding nonsense; however, it can’t replace human wisdom, empathy, or ethical decision-making. But let’s also recognize that burnout is a genuine risk to our field. Special educators leaving the profession create risks for students. Classrooms without qualified teachers pose risks to vulnerable populations. If an ethical, teacher-controlled AI pro- tocol helps dedicated professionals remain in work they love longer, that future is worth building carefully and intentionally. RECLAIMING WHAT BROUGHT US HERE When I began experimenting with AI last spring, my only goal was survival. What I discovered was transformation. Today, I no longer count days until retirement. I’m back to counting students, celebrating victories, and cherishing con- nections. AI didn’t restore my love of teaching—my students did that, especially those in my “FUN”ctional math class. AI sim- ply cleared enough space in my overwhelmed schedule to see students clearly again. If KAIIT can accomplish that for one burned-out teacher, it can help many others. I hope this article provides a starting point, a replicable structure, and permission to explore a tool that—used wisely and ethically—can lift the heaviest burden special educators carry. Our students deserve our best work. So do we.

BACK TO CONTENTS

PRODUCT INFORMATION

KAIIT (KAIRO ARTIFICIAL INTELLIGENCE IEP TOOL) Producer: Created by educator Tim Ranis (and Kairo) Cost: Free Description: KAIIT is not proprietary software but a copy- righted replicable prompt framework and workflow that any educator can implement with district-approved AI platforms. It carries no licensing fee, stores no student information, and relies entirely on teacher-controlled prompts and documents. The complete KAIIT master prompt and workflow documenta- tion are available at no cost to educators and are linked above.

CHATGPT Producer: OpenAI

Free Tier: Access to GPT-4o mini for basic experimentation ChatGPT Plus: $20/month for GPT-5.1 with file upload capabil- ity Enterprise: Variable pricing with enhanced privacy guarantees GOOGLE GEMINI Producer: Google DeepMind Free Tier: Access to Gemini 2.5 Flash Gemini Advanced: $19.99/month for Gemini 3 Pro Gemini for Education: Variable pricing with additional privacy guarantees

CLAUDE Producer: Anthropic Free Tier: Access to Claude Sonnet 4.5

Claude Pro: $20/month for Claude 3.5 Sonnet Enterprise: Variable pricing with security features

Note : KAIIT works with any AI platform. Most teachers can begin with free versions before determining whether paid sub- scriptions provide sufficient value for their workflow.

BACK TO CONTENTS

Page 1 Page 2 Page 3 Page 4 Page 5 Page 6 Page 7 Page 8

www.closingthegap.com

Made with FlippingBook Ebook Creator