The shared inbox collects everything.
Supplier invoices, requests for a quote, delivery notes, customer questions, and a fair amount that is none of those. Somebody opens each message and decides who should see it.
That person becomes the bottleneck. When they are on holiday the mailbox fills up, and a message that needed an answer the same day is found a week later underneath forty others.
One entry point, one classification step, then a separate path per document type. Each path ends in something actually happening.
A message and its attachments arrive in the monitored mailbox.
Each attachment is read and identified by type. The message body is read as context, not as the decision.
Into the invoice flow: fields extracted, totals checked, record handed to accounting.
→ BookedInto the quote flow: matched to the request it answers and stored alongside it.
→ FiledMatched against the order it belongs to and archived with that order.
→ MatchedTo the person or team responsible, with a short summary and the original attached.
→ AssignedWhat you get
The decision comes from what the document contains, so a file called scan_0043.pdf is no harder than one named properly.
A message with four attachments becomes four documents, each routed on its own merits.
You decide which type goes where and what happens at the end of each path. We wire it to the systems you already run.
Anything the assistant is not sure about waits for a person, with the reason attached. Nothing is silently filed.
Sender, subject and body are read alongside the attachment, so a reply in a thread lands with the thread it belongs to.
Where it arrived, what it was identified as, where it went and who touched it, kept per document.
If you already have flows, we route into them rather than replacing them.
Show us your inbox problem
Describe what lands in your shared address on an average day and who sorts it today. We will map the paths with you.