Knowing which client is actually ready to file
The problem
Clients upload documents to a portal throughout tax season. Someone has to open each one, work out what it is, decide whether the return can be prepared yet, and chase whatever is missing. In season that arrives faster than anyone can keep up with, and the question the firm actually needs answered — who is ready to file, and who is waiting on what — has no answer that is current.
The insight
The bottleneck is not reading the documents, it is knowing what is still missing
Extraction is the easy half, and it is where most automation stops — you end up with well-parsed documents and still no idea where anyone stands. The useful output is the gap: given what this client has uploaded and what their return requires, what is outstanding? That single question turns a pile of uploads into a per-client readiness state, and it is also what makes automated chasing safe to switch on. A system that asks for documents it already has trains clients to ignore it, so the follow-up only ever requests what is genuinely absent.
What we built
- Ingestion of client portal uploads, with each document classified by type on arrival rather than when a preparer opens it
- Extraction of the fields that matter to the return, with low-confidence results routed to a person instead of guessed
- A per-client gap analysis comparing what the return requires against what has actually arrived
- Automated client follow-up requesting only the documents genuinely outstanding, so the ask stays credible
- A dashboard showing every client's position in the tax-submission pipeline, current without anyone assembling it
- Process changes alongside the software — the firm's intake and review steps were reworked around the new signal, not bolted onto the old workflow
Have something with a domain problem underneath it?
Those are the ones worth talking about. Describe it and you will get a straight answer about what it would take.
Start a conversation