File Records Into the Right Series as They Arrive
Your retention schedule is written in record series: employment records, financial source documents, contracts, student records. Your shared drive stores a filename, a date and a size. Nothing in the second tells you the first, so a person has to decide, and they decide three years late from a filename.
The short answer
Classification is cheapest at the moment a document arrives, because that is when its context is still obvious. A workflow triggered by File Uploaded or Email Attachment uses Ask AI About the File to read each document and branch on which record series it belongs to, then Rename applies your naming and Move files it into that series folder. Anything the branch cannot place confidently goes to a review folder instead, which matters more here than elsewhere: a document filed into the wrong series inherits the wrong retention period, and nobody notices until disposition.
Steps this uses
Before and after
What actually changes
As they arrive
- Inbox/scan0043.pdf
- Inbox/signed copy.pdf
- Inbox/IMG_2291.jpg
- Inbox/notes from call.docx
After the workflow
- Contracts/2026-04-02-msa-northwind.pdf
- Employment/2026-03-18-offer-letter-r-mensah.pdf
- Financial/2026-02-27-receipt-dataflow.jpg
- Needs Review/notes from call.docx
Three documents filed into the series their contents put them in, and one the workflow would not guess at.
Setting it up
Described once, then it runs
This is the sentence. Send it to the builder and the steps below appear on a canvas, wired and named, for you to change before anything runs.
When a document arrives, read it and decide whether it is a contract, an employment record, a financial record or general correspondence. Rename it date-first and file it in the folder for that series. If it does not clearly fit one, put it in Needs Review instead.
Name the series your schedule uses
The branches in the Ask AI About the File step are authored, not fixed, so they should carry the names your retention schedule already uses rather than generic document types. If your schedule says "financial source documents", that is the branch, because the point is that a person reading the folder later can match it to a line in the schedule.
Give unrecognised documents somewhere to go
Add a branch for what does not fit and point it at a review folder. A misclassified record is worse than an unfiled one: an unfiled document is visibly waiting, while one filed into the wrong series looks handled and quietly carries the wrong retention period.
Read the date off the document, not the upload
Retention clocks usually start at an event rather than at file creation. The Read Details from File step pulls a date printed inside the document, such as a contract date or a tax year, so the name carries the date that actually matters.
Put approval where the stakes are
Ask a Person to Approve can sit in front of filing for the series where a mistake is expensive, and be left out of the ones where it is not. Approving everything is how a review step gets ignored.
Why classification at cleanup never works
A document arrives with its context attached: somebody knows what it is, who sent it and why, because they are looking at it. Three years later that context is gone and all that remains is a filename somebody typed in a hurry. Classifying at intake is not merely more convenient than classifying at cleanup; it is a different task with far better inputs, and it is the reason drives that classify on arrival stay legible while drives that plan to sort it out later never do.
What the workflow does not decide
It files documents into the series you named. It does not write your retention schedule, does not decide how long a series is kept, and does not act when a period lapses. Those are governed by law, your regulator and your own records policy. What classification buys you is that the question "which series is this" has an answer for every document on the drive, which is the precondition for any retention decision rather than a substitute for one.
Legal holds sit outside this entirely
No branch in a workflow knows that counsel opened a matter last Tuesday. Classification makes records findable, which helps when a preservation scope has to be drawn quickly, but a hold is a legal judgement made outside any document system and is not something a workflow can apply, detect or release.
Limits
What to know before you build it
It runs on new files, not your backlog
This workflow fires when a file arrives. It will not go back over the documents already sitting in the folder, so switching it on leaves the existing pile exactly where it is. Working through what is already there is a scheduled run instead, which is a different workflow.
Organize the files you already haveIt needs a connected account
This workflow reads files out of an email inbox, so it cannot run until that account is connected. Connecting one is on the Max plan, or on any team workspace. There is no version of this particular workflow that runs on the free plan, because the files it organizes never arrive in The Drive AI on their own.
FAQ
Common questions
How do you classify documents into record series automatically?
A workflow reads each arriving document with the Ask AI About the File step and branches on which series it belongs to, then renames and files it accordingly. The branches are authored by you, so they carry the series names from your own retention schedule rather than generic categories.
What happens to a document it cannot classify?
It goes to a review folder rather than a guessed series. This is deliberate. A document filed into the wrong series inherits the wrong retention period and may be disposed of early or kept long past its date without anyone noticing, which is a worse outcome than an unfiled document sitting visibly in a queue.
Does this apply my retention schedule?
No. It answers which series a document belongs to, which is the input a retention decision needs. Applying the schedule is a separate job, and disposal should pause for a person regardless. See the retention schedule workflow for how that part is arranged.
Can it classify the documents already on the drive?
Yes. The same workflow can run on a schedule over a folder you point it at rather than only on arrival, so a backlog is read and filed the same way new documents will be. The classification is weaker on old documents because their context is thinner, which is an argument for reviewing the results rather than for not doing it.
Classify it while somebody still knows what it is.
Every document read and filed into its series on arrival, and the ones that do not fit held back for review.
5 GB free · No credit card required