Post, scans, PDFs, photos and the forms people fill in. SwiftCase reads them and puts the details where they belong on the case, from the first morning, without anyone re-keying a thing. It does the typing. Your people keep the judgement.
What you get
Whatever arrives, the bits you need end up in the fields you would have filled in yourself.
Reference numbers, names, dates, addresses, amounts. The fields you would have filled in from the front of the file are in the case before anyone opens it.
Attachments on incoming email can be read the moment they land. By the time someone picks the case up, the detail is already there waiting.
A photographed form, a scanned letter, a picture of a site, a plate or a meter. If a person could read it, so can this.
It is set up around your own forms and your own vocabulary. A site reference and a patient number are different things, and it knows which one you mean.
Everything it enters sits on the case like any other field. Your team can see it, check it and change it in a click.
Configured to your paperwork as part of implementation, so it is reading your documents from the first morning, not after a training period.
What it reads
The paperwork that arrives at a working firm is rarely tidy. A photographed form, a scanned letter with a coffee ring on it, a forty page PDF where the one number you need is on page thirty-one. That is the paperwork it is built for.
And because it is set up around your own fields, it is reading for the things you would look for, not a generic list of what a document might contain.
Built for work that has to stand up
Extraction is there to get the detail in front of the right person faster, not to decide anything on its own. The person who signs the report is still the person who made the call.
What it entered is on the case where your team already works, not in a separate queue they have to remember to check.
Any value it fills in can be changed like any other field. It never has the last word.
The fields and the vocabulary are yours, so it is reading for what you would look for, in the terms you use.
On every plan
It is useful on the first morning, and it improves just from your team doing their jobs.
Day one
It reads your paperwork and fills in your fields from the first case, before there is any history for it to learn from.
As you work
Every correction someone makes and every case that closes adds to the record. Nobody has to feed it.
When it earns it
Where your own history can do a better job than a general model, a model built on that history takes over, checked against what really happened first.
Quick answers about document extraction.
PDFs, Word and RTF documents, scanned pages, photographs and attachments on incoming email. The forms your own clients complete are read the same way. If a person could read it, it can be read.
No. It is set up to your paperwork and your fields during implementation and works from the first case. As your record builds, models trained on your own history can take over the parts where that earns its keep, so it gets better the longer you use it without a project to make it happen.
Everything it enters sits on the case as an ordinary field, visible to the person handling it and changeable in a click. It is there to do the typing, not to make the call. The person who signs the report is still the person who made the decision.
Yes. It is configured around your fields and your vocabulary rather than a generic set. A site reference, a patient number, a policy number and a job number are different things, and it is set up to know which is which for your firm.
Yes, from the entry tier. Higher tiers add reading attachments straight off incoming email, across more mailboxes.
Bring a real file to the demo and see what your own paperwork looks like once it reads itself.
