ABET SELF-STUDY REPORT AND AI
Can you write your ABET self-study report with an AI chat?
Yes, for parts of it. A general-purpose AI chat is good at turning notes into a clear paragraph, and many programs already use one. Where it stops is the record behind the paragraph. A chat doesn't keep your evidence, your course to outcome mapping or your continuous improvement cycles from one year to the next. And when the evidence isn't there, it fills the gap with something that sounds right. Digital Accreditation Lab (dal.ai) uses AI too. The difference isn't the AI. It's the system of record around it, so every claim in the draft points to the evidence behind it.
45 minutes with example documents. Nothing to prepare, no account, no card.
An AI chat and a shared drive
Loose files
In drives and inboxes
Pasted into a chat
By one person
A paragraph
No sources shown
The work walks out
When the person leaves
dal.ai
Your evidence
From your team
The record
Mappings and CI loops
A cited draft
Gaps named
Your sign-off
The record stays
The record carries into the next cycle.
Where the evidence lives
- An AI chat and a shared drive
- In a drive, apart from the chat
- dal.ai
- In one record, classified and accepted by you
Getting files from faculty
- An AI chat and a shared drive
- Email requests and attachments
- dal.ai
- A link to upload exactly what you asked for, no account needed
The source behind each claim
- An AI chat and a shared drive
- You check it yourself
- dal.ai
- Cited paragraph by paragraph
When evidence is missing
- An AI chat and a shared drive
- Fills the gap
- dal.ai
- Names what is missing
Criterion 4 across years
- An AI chat and a shared drive
- Rebuilt from memory
- dal.ai
- Kept in four stages, closed when you verify
ABET's edition
- An AI chat and a shared drive
- Whatever it was trained on
- dal.ai
- Your commission's criteria and questionnaire
Who works in it
- An AI chat and a shared drive
- One person at a time
- dal.ai
- Program Director, Program Coordinator and Faculty, each in their own role
When someone leaves
- An AI chat and a shared drive
- The work leaves too
- dal.ai
- The next person opens the same record
| An AI chat and a shared drive | dal.ai | |
|---|---|---|
| Where the evidence lives | In a drive, apart from the chat | In one record, classified and accepted by you |
| Getting files from faculty | Email requests and attachments | A link to upload exactly what you asked for, no account needed |
| The source behind each claim | You check it yourself | Cited paragraph by paragraph |
| When evidence is missing | Fills the gap | Names what is missing |
| Criterion 4 across years | Rebuilt from memory | Kept in four stages, closed when you verify |
| ABET's edition | Whatever it was trained on | Your commission's criteria and questionnaire |
| Who works in it | One person at a time | Program Director, Program Coordinator and Faculty, each in their own role |
| When someone leaves | The work leaves too | The next person opens the same record |
What an AI chat does well
Be fair to the tool: it turns rough notes into readable prose, summarizes a long report, rewords a section that isn't landing, and explains how a criterion is usually addressed. If your evidence is already organized and one person owns the report, a chat can genuinely help.
Where the do-it-yourself method breaks
- It forgets. Each conversation starts over. The evidence, the decisions and the prompts live in one person's account, and when that person changes roles, the record leaves with them.
- It fills gaps. Ask for a Criterion 4 narrative with thin input and you get a confident paragraph. Finding what was invented falls to whoever reviews it.
- It can't follow a loop across years. Criterion 4 asks for assessment, evaluation, action and re-assessment over a six-year cycle. A chat has no way to know that an action from two years ago still needs to be re-checked.
- It doesn't know your edition. ABET publishes a new self-study questionnaire for each commission every year. A chat works from what it was trained on or what you paste in.
- It has one seat. Accreditation is team work. Faculty supply files, a coordinator organizes them, a program director decides. A chat window belongs to one person.
What dal.ai keeps that a chat can't
dal.ai is the system of record for ABET accreditation. You upload evidence once, or send faculty a link to upload exactly the files you asked for, with no account to set up. dal.ai proposes which criteria and student outcomes each file supports and quotes the passage behind it, and you accept or correct each one. Courses are mapped to student outcomes from their syllabi. Each continuous improvement cycle is kept in four stages: assessed, evaluated, action taken, verified. A loop counts as closed only once you accept that it was verified.
When it's time to write, dal.ai drafts each section from the evidence you have accepted and your commission's ABET criteria and questionnaire. It cites the source behind every claim and names what is missing instead of filling the gap. Nothing reaches ABET without your sign-off. And when your coordinator moves on, the next person opens the same record.
Built for the whole team
Accreditation is never one person's job. Faculty hold the syllabi, the rubric results and the student work. A Program Coordinator gathers and organizes them. A Program Director makes the calls. With an AI chat, all of that funnels through whoever holds the account, and the rest of the team works from email attachments and a shared drive.
dal.ai gives each role a place in one shared record. When you need specific files from a specific person, you send them a link to upload exactly those files, with no account to set up. The Program Coordinator organizes what comes in, the Program Director reviews and accepts, and everyone works from the same evidence and the same draft. When someone changes roles, the record stays with the program.
Faculty
Send files by link, no account
Program Coordinator
Organizes the evidence
Program Director
Reviews and accepts
One shared record
Evidence, mappings, CI loops and drafts
The next person
Opens the same record
One continuous improvement loop in dal.ai, from assessment report to the Criterion 4 section.
Narrated with an AI voice.
Transcript
Continuous improvement is common sense. Find what's falling short, work out why, change something, and check that it worked. The hard part is the record. Criterion 4 asks for years of that work, and most of us can't recall what the committee decided last semester.
dal.ai is the system of record. Each file goes in when it's made: the report when it's done, the minutes after the meeting. Then the self-study doesn't depend on anyone's memory. For this demo, we'll add a whole loop at once. First, you tell dal.ai how your academic year runs.
Upload the assessment report. You choose which academic year its results are about, and dal.ai checks your choice against the file. Start the cycle, and the results are there, each with its source. Add the committee's minutes, and dal.ai drafts the evaluation: why two indicators fell short.
You review it and accept it. It shows what it used, and what it left out: an item about another outcome. Then the action: three changes from Fall 2025, and a plan to reassess. A year later, the reassessment and the closing minutes. dal.ai compares every indicator before and after, and records the committee's vote to close the loop.
From that record, dal.ai drafts the Criterion 4 section. It flags what the record doesn't cover yet, and the section can't be marked complete until you deal with that. Every claim cites its file. Click one, and you're on the file it came from. The loop is closed.
SO 6 goes back into the rotation, and the record is already there.
What doing it yourself really costs
An AI chat subscription costs little. The expensive part is the record: someone spending the months before a visit finding last cycle's rubric results, working out which action answered which finding, and checking every sentence against a file. dal.ai keeps that work from piling up, because the record is kept as you go.
A note on ABET's AI policy
ABET's published AI policy allows AI to gather and synthesize information, support the analysis of assessment data, and help organize supporting materials. Qualified people must verify everything, AI must not alter or misrepresent evidence, and programs should acknowledge their use of AI. That holds whichever tool you use. With a chat, staying within it depends on each person's habits. dal.ai builds it into the workflow: it drafts without changing your evidence, records who accepted what, and leaves the judgment with your people.
dal.ai is independent and is not affiliated with or endorsed by ABET.
Frequently asked questions
Can I just use ChatGPT to write my ABET self-study report?
You can use ChatGPT or any general AI chat to help draft parts of it, within ABET's AI policy. It works well for turning your notes into clear prose. What it doesn't do is keep your evidence, your outcome mapping and your continuous improvement record, or show the source behind each claim. Checking every sentence and keeping the record stay with you.
Isn't dal.ai just AI too?
dal.ai uses AI to read evidence and draft sections. The difference is what surrounds it. It works only from evidence you have accepted and from your commission's ABET criteria and questionnaire. It cites the source behind every claim, and it keeps the record in a shared workspace. The AI proposes. You decide.
Do faculty have to learn another system?
No. When you need files from a faculty member, you send them a link to upload exactly what you asked for, with no account to set up. Faculty you invite to the program can also sign in and add files themselves.
Can we keep using our AI chat alongside dal.ai?
Yes. Keep a chat open for wording or brainstorming if it helps. dal.ai is where the evidence, the decisions and the cited draft live.
Is it safe to paste student work into an AI chat?
It depends on the account and your institution's rules. A personal account isn't covered by an agreement with your institution. dal.ai processes documents through Anthropic, our AI model provider, under commercial terms that do not allow your content to be used to train its models. dal.ai also signs a FERPA agreement with your institution and acts as a school official for any student information you upload.
Does dal.ai write the self-study for me?
No. It drafts sections from your accepted evidence and names what is missing. Your people review, edit, approve and attest.
ChatGPT is a trademark of OpenAI. Digital Accreditation Lab is not affiliated with OpenAI.
See Digital Accreditation Lab before you decide anything
Start with a 45-minute session. We use example documents, so there is nothing to prepare, and you see dal.ai classify files, map a course to student outcomes, and draft a cited self-study section with the gaps named. If you want more, try it privately for 30 days on one criterion with a few colleagues you trust, at no cost within the AI usage limits in our fair use terms. Every step is optional, and nothing reaches ABET without your sign-off.
Book a 45-minute sessionDigital Accreditation Lab (dal.ai) is an independent product of FindBalans LLC. It is not affiliated with, endorsed by, or certified by ABET.