The Status Update You Wrote by Reading Six Screens
Answering 'where's my return up to?' shouldn't mean digging through five systems. Here's how AI and a single record turn a five-minute hunt into a one-line answer.
A client rings on a Tuesday afternoon. The question is simple: "Where's my company tax return up to?" The answer, it turns out, is not. You check the workflow board to see what stage the job is on. You open your emails to remember what you last asked for. You look in the portal to see whether they've uploaded the missing bank statement. You glance at the lodgment program to confirm the due date. You check the invoicing system to see whether the deposit came through. Then you cobble together a sentence that sounds confident and hang up.
Five minutes of hunting for a thirty-second reply. And you'll do it again tomorrow for the next client who asks.
Why the simple question is so hard to answer
The status of a job is never in one place. It's spread across the work item, the correspondence, the documents, the compliance calendar and the ledger — and in most firms those live in separate tools that don't talk to each other. To answer "where are we up to?" you have to be the integration layer. You hold the full picture in your head, or you reassemble it from six screens every single time.
This is the hidden tax of fragmented systems. It rarely shows up on a timesheet because nobody codes "looked things up to answer a client" as billable work. But across a practice, it's hours a week — and it's worse when the person answering isn't the one who did the work. A staff member covering for a colleague on leave can't reconstruct context from a spreadsheet and an inbox. So the client waits, or gets a vague answer, or gets bounced to voicemail.
The fix isn't more effort — it's one record
The reason status is hard to summarise is that the underlying information is scattered. Bring it together and the summary becomes trivial. When your accounting client management software holds the client record, the work items, the documents, the deadlines, the engagement and the invoice against the same client — in one system — the picture assembles itself.
In Finye, a client isn't a name repeated across five tools. It's a single record. The work item sits on a board with a real stage. The portal requests and uploads hang off that same client. The compliance deadlines the firm owes are tracked in the deadline register. The engagement letter and its signature status are there. The invoice and payment are there too, because invoicing lives with the work rather than in a separate client accounting software silo. When everything is attached to one record, answering "where are we up to?" stops being a research project.
Where AI earns its place
Consolidating the data is half the job. The other half is turning it into a sentence a human can read out loud. That's exactly what AI is good at — not replacing your judgement, but reading the current state and drafting the plain-English summary.
Instead of you scanning six screens and composing the update, the system already knows:
- the job is at Preparation, not yet Review
- you're waiting on one item — last quarter's bank statement — requested through the portal four days ago
- the deposit invoice was paid on the 3rd
- the engagement letter was signed in July
- the return is due next month, with buffer
AI can turn that into: "Your company return is in preparation. We're waiting on your Q4 bank statement, which we requested on the 12th — once that's in, we'll move to review. You're well ahead of the due date." One line, accurate, ready to send or read. You've done no digging. The facts came from the record; the phrasing came from the AI.
This is the difference between AI and automation
It's worth being precise about the roles. Automation moves things without you: it advances a job when a document lands, it fires the next portal request, it flags a deadline that's approaching. AI interprets and communicates: it reads the assembled state and writes the human-facing summary. Neither is magic, and neither works if the data is scattered. Both work beautifully when there's a single source of truth to draw from.
What this changes in practice
When status is instant, a few things shift.
Anyone can answer. The receptionist, the junior, the partner covering for someone on leave — they all see the same current picture and can give the same confident answer. Knowledge stops living in one person's head.
Clients ask less. A lot of "where are we up to?" calls happen because the client is anxious and can't see progress. A portal that shows them their own outstanding items, backed by proactive updates, heads off the call entirely. The best status update is the one the client didn't need to request.
You stop paying the reassembly tax. The minutes you spent reconstructing context go back into the work. Multiply that across every enquiry, every day, and it's real capacity.
What to look for
This isn't about buying a chatbot and hoping. If you're evaluating account practice management software with automation and AI in mind, the order of operations matters:
- One record first. Client, work, documents, deadlines, engagement and invoicing on the same client — not stitched together from a ledger, a tax return software tool, an inbox and a spreadsheet. AI is only as good as the data it can see.
- Automation that reflects reality. Jobs that advance on real triggers — a signed letter, an uploaded file, a received payment — so the status is current without anyone updating it by hand.
- AI that drafts, not decides. Summaries and draft replies you review and send, grounded in the actual record rather than guessing.
The question "where's my return up to?" will never stop coming. But it should never again take six screens and five minutes to answer. When your client accounting work, correspondence and compliance all sit on one record, the answer is already written — you just have to read it out.