Included on every plan
Read the packet. Draft the answer. You decide.
Hand TRACE a technician's certification packet or a client's procedures and get back structured, cited draft findings — extracted records, checklist assessments, answers about your own portfolio. Every one of them is a suggestion you review before it becomes a record.
What it does
AI Assist, in detail.
Turn a packet into records
Upload the pile a new technician arrives with — certificates, training records, OJT logs, vision exams — and get back their details, each certification, and the requirements behind it, with the hours, scores and dates already filled in. Every value cites the document it came from and a quote showing where.
A first pass at the checklist
Point it at a client's procedures alongside an AC7114 checklist and it works the criteria section by section: Conform, Nonconformance, or Not Applicable, with the client's own procedure reference, the reasoning, and the quote it relied on. You get a populated draft to argue with instead of a blank checklist.
Ask about your own portfolio
"Which UT technicians are due in the next 60 days?" — asked in plain English, answered from your live records. It looks things up rather than guessing, and it only ever sees what your own login is permitted to see.
It drafts, it never writes
Nothing reaches a record until you accept it. Extracted certifications land in a review screen where you check each one; checklist assessments populate a draft you edit; the assistant can read your data but cannot create, change, or delete anything at all.
Consultant-side only
Client and technician logins have no access to any of this. That's deliberate: these outputs are drafts pending your review, and handing an unreviewed verdict to the company whose technicians are being judged skips the very step that makes it trustworthy.
Metered in credits, not tokens
Included on every plan, Free included. A question costs one credit, a checklist run costs more because it's a much bigger job, and your allowance resets on the 1st. Your Account page shows what's been used and what's left, so the bill is never a surprise.
How it works
Day to day.
- 1
It's already on
AI comes with every plan, including Free. Nothing is sent anywhere until you actually ask for something.
- 2
Give it the documents
Upload a technician's packet, or pick the client procedures to assess a checklist against. It reads what you supply — not the open internet.
- 3
Review what comes back
Each finding shows its source document, a supporting quote, and a confidence level, so you can check the ones that matter rather than re-reading everything.
- 4
Accept what's right
Take the rows you agree with into real records and discard the rest. What you rejected is never written.
Where we draw the line on AI
This drafts, it doesn't decide. Two limits are deliberate and worth stating plainly. First, it will not fill in required training or experience hours from NAS 410 or SNT-TC-1A, even though it could produce plausible-looking numbers: under both standards your employer's written practice is the governing document, and it routinely sets thresholds above the published minimum — a confidently wrong "required hours" figure is the most damaging thing this feature could produce. It reports what your documents say and flags what's missing. Second, it can misread a document, and a cited quote is there so you can check it in seconds rather than take it on trust. Nothing it produces is a record until a consultant accepts it, and no certification decision is ever made by the model.
Want to see it on your own program?
Tell us how your NDT program runs today and we'll show you how this fits it.