Your tax portal is live. Why is the team still chasing documents by email?

Give each request one status the team can trust

Choose the record your team will use to decide what to request next and whether work can begin. It may already be in your client portal or practice-management system. A second spreadsheet won’t solve the problem if staff still have to update both places to keep a request current.

Make the collection owner responsible for that record, with a named backup. Each requested item needs a client and return ID, reporting period, request ID, status, evidence link, next action and owner. The period matters because a statement from the wrong year can look like a response while leaving the current request unresolved.

Separate receipt from acceptance. “Received, awaiting check” means the file has arrived. “Accepted for this request” means the assigned person checked it against the firm’s requirement. “Waiting on client,” “waiting on third party” and “access problem” keep different causes visible, so the coordinator doesn’t send the same reminder for all three.

For the example, the coordinator matches the emailed statement to request R-01, checks the client and period, and stores or links the file in the firm’s approved document system. The coordinator updates R-01 and stops its reminder. Request R-02 remains open for the missing document, which is why accepting the statement doesn’t make the whole return ready.

The tax manager defines which items are required and who can approve an exception. If limited work can proceed, record exactly which work is allowed, who approved it and what remains held. A general “ready” flag would hide that boundary from the preparer.

You now have a way to remove repeated chasing without assuming every delay is under your control. A document that hasn’t been issued still needs an owner and a next check date, but another reminder can’t make it exist. If missing inputs aren’t the cause of your backlog, the tax-review queue guide helps separate readiness problems from a shortage of qualified review capacity.

* Eight-week delivery, or twelve weeks when substantial data foundations are required, as agreed before work begins. If a covered stage fails its agreed benchmarks, we provide up to four additional weeks of remedy work at no charge, then waive or refund that stage’s fee if it still fails. Successfully delivered stages remain payable. Agreed client-caused delays pause the clock. Delivery guarantee terms.

Move the open work before you turn off the old queue

The new record can be correct for new requests and still miss everything already sitting in an inbox. Before switching one return type, have the coordinator reconcile its open email threads and existing tracker against the new request list at a stated cutoff time.

Match each item by client, return, period and requested document. Bring across the latest reply, promised delivery date, owner and unresolved question. Merge duplicate requests into one active item while retaining the source references. Don’t send a new request just because an old one has moved systems, since that asks the client to solve your migration problem.

Check in both directions. Every old open item should map to a new item or a documented closure. Every new item should have a source and an accountable owner. A matching total isn’t enough: two lists can both contain ten requests while omitting different clients.

Keep the old tracker available as a read-only reference after the switch. Tell staff which record now owns status and how to route replies from old threads. The coordinator still monitors the agreed communication channels, but records their effect in the new queue. Retiring duplicate tracking doesn’t mean ignoring a client’s reply or deleting the evidence behind the work.

Set the switch time, responsible owner and recovery route before the first live request. If the team loses access or can’t reconcile an important item, pause the affected handoff and route it to that owner. Use the firm’s agreed temporary log during an outage, then reconcile those entries before normal processing resumes. Two independently edited queues would recreate the uncertainty you’re trying to remove.

Keep the late reply and the access problem in the same process

Now change the example. The client can’t use the portal and replies to last week’s email. The coordinator records the access problem on the existing request, assigns someone to resolve it and gives the client the firm’s approved alternative way to supply the document. Staff can help the client without opening a private tracking process that no one else can see.

The bank statement arrives through that approved route, but the other document won’t be available until Friday. Record Friday as the next check date on R-02, together with the source of that expectation. If Friday passes without the document, the collection owner follows up or escalates under the firm’s rule. A due date without that owner would only make the delay visible.

Use the structure below as one request row in your current system. The table turns the row vertically so the fields are easy to copy. The worked column shows R-02 after the statement has been accepted for R-01. The coordinator fills it in, and the tax manager owns readiness exceptions.

This lets a replacement coordinator pick up the work without reconstructing the inbox. It also gives the preparer a specific reason for the hold. If the client supplies the document on Thursday, the owner updates the same request and cancels the Friday chase after acceptance. The scheduled action follows the current evidence rather than becoming another message sent automatically.

Test the handover before adding more automation

Check your current tool first. Caseware’s client-request guide describes portal requests with statuses and due dates, responses and returned items. Use the equivalent features in your own setup before buying or building another collection layer. Product features still need a firm rule for ownership and acceptance.

For one return type, have the collection lead walk representative requests through the new process while the existing process remains authoritative during the trial. Include a normal upload, an old-thread reply, a duplicate file, a wrong-period document, an access failure and a delayed third-party item. For each, confirm that the correct request receives the evidence, the right person owns the next action and the readiness decision follows the approved rule.

A test fails if an unresolved required item disappears, a duplicate reminder continues after acceptance or a preparer starts work outside an approved exception. Repair the failed route and retest it before switching. Once the trial passes and the open-item reconciliation is complete, the collection lead records acceptance and the switch time. That’s when the old tracker becomes read-only.

If matching incoming replies is still consuming substantial time, test AI as an assistant to the coordinator. Give it only permitted correspondence, request IDs and the firm’s matching rules. Ask it to propose the matching request, received item, period and source passage, with “unresolved” when the evidence conflicts. Keep its proposal separate from the authoritative status until a person accepts it. A plausible match to the wrong client would otherwise stop the right reminder and leave a real request unattended.

Compare those proposals with coordinator-labelled examples before using them on live work. Count wrong-client matches, missed replies, duplicate matches and the time spent checking and correcting the suggestions. Continue with manual matching or current-tool rules if the assistant adds more work than it removes. A model isn’t needed to enforce a fixed required-item list, and it doesn’t approve tax judgments or readiness exceptions.

After switching, compare duplicate contacts, unowned requests, preparation restarts and total collection effort for comparable return types over the agreed observation window. Include exception handling and checking time, and keep third-party waiting separate from staff effort. That shows whether the change removed work or merely moved it to the coordinator. The capacity and pricing guide explains how to follow any released time into completed work or an expense the firm can actually stop paying.

Make one cutover decision the collection owner can explain

Take one return type with duplicate tracking and ask its collection owner to complete this sign-off. Use the request record above to support each answer.

The payoff is a clear decision: switch this return type, or name the exact failed check that prevents it. When the checks pass, stop maintaining the duplicate status record at the agreed time. You’ll still have a route for late replies and missing documents, but the team will know where the next action lives.

Removing duplicate tracking can give the tax team more time for completed client work. Whether that is the best investment for your firm’s next stage of growth depends on what most limits the business goal. Use the cutover evidence to investigate that question alongside review capacity, pricing and other demands on the team.

* Eight-week delivery, or twelve weeks when substantial data foundations are required, as agreed before work begins. If a covered stage fails its agreed benchmarks, we provide up to four additional weeks of remedy work at no charge, then waive or refund that stage’s fee if it still fails. Successfully delivered stages remain payable. Agreed client-caused delays pause the clock. Delivery guarantee terms.