Soraban Live Demo: Join our weekly webinar
Register Now
AI-native workpaper software turns source documents into structured, reviewable work. Each value stays tied to its original support, and preparers and reviewers remain in control of the decisions that affect the return.
AI features now appear across accounting products. Some extract fields, answer questions, draft notes, or summarize files. Those functions can be useful, but they don’t necessarily change how a tax file moves through preparation and review.
Soraban is built for that execution gap. It uses accountant-trained AI to move work through intake, workpaper preparation, data entry, and delivery while preserving visibility, review, and control. The practical test is straightforward: What work does the system complete, what can your team inspect, and how soon can usable documents move forward?
In tax workpaper software, AI-native means the AI is built into how preparation gets done. It works from source documents, return context, firm structure, and reviewer input to complete repeatable tasks inside the engagement.
Summarizing or extracting information from one document can be useful, but workpaper preparation depends on how that document fits into the rest of the return.
The system needs to retain details such as the taxpayer, tax year, document type, binder location, related workpaper, missing information, preparer activity, reviewer decisions, and approved annotations. As new documents arrive, it can use that context instead of treating every upload as a separate task.
Summarizing a K-1 or answering a question about it is only one step. The more important question is what happens next.
Tax workpaper software can recognize the form, confirm the taxpayer and year, place it in the correct section, populate the related work, tie values to source, and flag missing or conflicting information. It may also draft a reviewer note for the tax team to approve or edit.
That keeps the work inside the return file instead of asking someone to copy an AI response elsewhere and rebuild the context around it.
Workpaper quality takes shape before prep. How documents are collected, matched, tracked, and organized determines how much cleanup remains when the preparer opens the file. Intake is the exact place where the phrase “garbage in, garbage out” can create bottlenecks and inaccuracies.
Many firms rely on organizers, portals, email, spreadsheets, and document tools. Those processes can still leave manual decisions when submissions arrive in stages, file names are unclear, pages are missing, or documents come through several channels.
Collect builds requests and checklists, tracks what’s missing, sends automated reminders, and organizes files as they arrive. This gives preparers a cleaner, prep-ready file without expecting clients to submit one complete, perfectly labeled package.
Prep doesn’t have to wait for one complete batch. Usable documents can move forward while unresolved items remain visible. In practice, the sequence looks like this:
This keeps one late document from stalling unrelated work or forcing the team to rebuild the file when another item arrives. The affected section still needs complete support and review before it can be finalized.
Once documents enter the workspace, they need to land in a structure the tax team can prepare and review. A sorted file list doesn’t show how each source connects to the return, related calculations, open questions, or reviewer decisions.
Tax teams need a file structure that shows where documents belong, how amounts roll through the return, and where reviewers can find the supporting work. Keeping each layer focused on its own job makes the file easier to prepare, follow, and review.
Prepare uses that structure to place documents and supporting work in the right part of the file as they arrive.
The system needs the firm’s structure before information can land in the right place. In Soraban, the questionnaire sets up the workspace and its binder, leadsheets, and workpapers.
The setup reflects the firm’s section structure, naming rules, templates, expected sources, and preparation responsibilities. The AI follows those conventions as it processes documents.
Prior-year information can guide the questionnaire and expected-document list, but current-year facts and support control the work. A previous treatment or conclusion shouldn’t carry forward without current support.
Tax files rarely arrive as one clean, final package. Work that’s ready can keep moving while exceptions remain visible and easy to resolve.
Each unresolved item should show its status, owner, date, affected workpaper, and next step. It may involve a missing form, unanswered question, incomplete page, low-confidence match, or value that conflicts with another source.
Open items remain tied to the affected work instead of getting buried in email threads or general notes. The team can resolve them without holding up unrelated sections.
When a corrected or replacement document arrives, the system needs to flag changed values, identify the affected section, and show what requires another look.
The earlier source and reviewer annotations remain intact, so the new form doesn’t overwrite prior work or hide why a number changed. AI surfaces the difference, while the preparer or reviewer decides how it affects the return.
A review-ready workpaper gives the tax team enough context to understand the source, follow the calculation or treatment, see what remains open, and confirm how the file reached its current state. Review depends on that visibility.
A populated value should lead back to the supporting document, page, and relevant location. Reviewers shouldn’t have to search the binder to find where a number came from or confirm that the source belongs to the correct taxpayer and tax year.
The trail may include document references, page links, cross-references, tie-outs, and flags for duplicate, incomplete, unreadable, or conflicting information. Confidence indicators can direct attention, but they don’t replace the source.
Clear traceability gives reviewers a direct way to verify the work and preparers a clear place to correct a mismatch.
Preparer notes, open questions, reviewer edits, cleared items, approved treatments, rejected suggestions, and signoff status remain connected to the relevant workpaper.
Prepare can draft reviewer notes for the tax team to approve or edit. Accountant annotations and changes stay intact, so another round of processing doesn’t erase the reasoning already added to the file.
The review trail shows what’s complete, what changed, what was resolved, and what still needs attention.
AI can complete repeatable preparation tasks, but it can’t take responsibility for the tax treatment or final conclusion. Preparers and reviewers still need to assess the evidence, apply firm policy, resolve exceptions, and approve the work.
The split is straightforward: AI handles repeatable preparation work, while the tax team makes the decisions that affect the return.
AI-supported execution
Professional review and judgment
This keeps routine work moving without blurring responsibility. Preparers review populated work and resolve follow-up, while reviewers focus on treatment, consistency, and signoff. Source support, changes, and approvals remain visible so each decision can be checked in context.
Review-ready workpapers need to keep the return moving. Once the work is approved, the same information should carry into tax data entry and delivery without extra file handling, rekeying, or status chasing.
A connected tax workflow depends on each stage passing organized, usable work to the next. Soraban supports that movement across Intake → Prep → Data Entry → Delivery, with each product handling a defined part of the process.
Collect brings documents in and organizes them. Prepare turns those documents into reviewable binders, leadsheets, and workpapers. Connect moves reviewed data into the firm’s tax software, and Deliver handles the signatures, payments, reminders, and closeout needed to finish the return.
Because each stage passes usable work forward, the team spends less time reorganizing files, retracing source documents, or checking what still needs attention.
Workpaper software sits between the systems that organize the engagement and the software that calculates the return. Practice management organizes work. Tax software calculates returns. Soraban moves work through the firm.
In practice, that means fitting into the firm’s current stack, preserving established review steps, and reducing manual handoffs without forcing teams to replace the systems they rely on.
Before adding another platform, ask a few practical questions:
At the handoff into the tax program, the evaluation becomes specific: what transfers, what remains available, and what staff still need to touch. Soraban supports workflows with CCH Axcess, Lacerte, UltraTax, and Drake, so firms can test that fit against their current review process.
Once workpaper software handles tax documents and reviewed data, security and approval controls become part of the tax process. Firms need to know who can access the work, how long information is retained, and how AI output is reviewed before it moves forward.
A security review needs to cover the controls that affect daily work:
Soraban is SOC 2 Type II compliant and uses encryption, access controls, logging, audit trails, and retention safeguards. Client documents, extracted data, and firm data aren’t used to train third-party commercial models.
The firm still decides who can access a return, approve work, retain information, and remove access as responsibilities change.
AI output needs to remain under the control of the preparers and reviewers responsible for the return. They need to see the source, understand what changed, and decide whether the work is approved, revised, rejected, or overridden.
Accountant annotations and edits remain with the file throughout review. This keeps the decision trail intact before reviewed data moves into the tax program.
The value of workflow automation shows up in how returns move through the firm, not only in how quickly software reads a document or populates a field.
Useful measures include time from first document to prep start, manual classification, manually created workpapers, open-item aging, review cycles, data-entry effort, time spent locating source support, late-document rework, adoption by role, and completed returns during deadline periods.
Together, those measures show where repeated effort is declining. Admins may spend less time checking status, preparers may start with better-organized files, and reviewers may spend more time on exceptions. Any capacity gain depends on the firm’s current process, staff adoption, review standards, and how broadly the platform is used.
A pilot needs to reflect the work the firm actually handles, including the exceptions that create extra review and follow-up. A polished demonstration may not show what happens when documents are incomplete, corrected, hard to read, or assigned to the wrong return.
Use a pilot set that reflects the firm’s return mix and includes a few realistic exceptions:
Ask the vendor to show how the system handles a source mismatch, reviewer correction, unresolved open item, and failed handoff. The test should follow the work through preparation, review, and downstream movement rather than stop at extraction.
Before rollout, assign responsibility for workspace setup, templates and firm conventions, intake monitoring, exception handling, preparer review, reviewer approval, access management, vendor escalation, and success measurement.
Admins, preparers, reviewers, managers, and partners need role-specific training and clear acceptance criteria. The rollout also needs to define where AI is used, what still requires review, and what staff do when the system is uncertain. A pilot can reveal fit, but it can’t guarantee busy-season results.
Workpaper software is AI-native when AI performs tax-specific tasks inside the return file, keeps context as documents arrive, and produces source-linked work for review. What matters is how the system carries out the work, not whether it includes a chatbot or a stand-alone extraction feature.
OCR reads text and values from documents. AI-native workpaper software carries that information into the tax workflow by identifying the source, placing it in the right file structure, populating related work, tracking exceptions, and keeping each result tied to evidence the tax team can review.
AI can handle repeatable work such as document classification, source matching, workpaper population, cross-references, initial tie-outs, open-item detection, and draft notes. Tax treatment, unusual transactions, estimates, elections, exception resolution, and final approval remain with the preparers and reviewers responsible for the return.
AI can prepare repeatable parts of a tax workpaper, but the resulting work still requires accountant review. The tax team assesses the support, resolves conflicts, applies firm policy, evaluates the treatment, and approves the work before it becomes part of the return.
Yes. Preparation can begin for sections supported by usable documents while missing items remain visible and unrelated work continues moving. The affected section still needs the required information, any necessary follow-up, and professional review before the work can be finalized.
A late or replacement document should trigger a review of the affected workpapers. The system needs to flag changed values, preserve the earlier source, and keep reviewer notes intact so the preparer or reviewer can decide how the new information changes the work and the return.
Reviewers need the source document, page reference, populated value, open items, notes, changes, overrides, and approval status. They should be able to trace each result to its support, see what changed, and understand how the work reached its current state without reconstructing the file.
No. Practice management organizes the engagement, tax software calculates the return, and workpaper automation moves preparation work between intake and calculation. It fits alongside those systems to reduce manual handoffs while preserving the tools, conventions, and review steps the firm already relies on.
Firms should look for SOC 2 Type II, encryption in transit and at rest, multi-factor authentication, role-based access, activity logging, retention controls, customer data ownership, and limits on model training. They also need clear rules for who can approve, change, export, retain, or delete information.
Track the work the platform is expected to change: setup time, manual touches, open-item aging, review cycles, data-entry effort, source-search time, late-document rework, and adoption by role. ROI depends on return volume, implementation, review standards, staff adoption, and how broadly the software is used.
A useful workpaper platform should begin working as source documents arrive, organize the file around the firm’s structure, preserve source traceability, keep open items visible, and give preparers and reviewers control over changes. The result is less repeated setup, stronger review context, and a cleaner handoff into the systems that calculate and complete the return.
Soraban carries that structure through the rest of the tax workflow, keeping documents, review context, approved data, and client follow-up connected across Intake → Prep → Data Entry → Delivery. The result is a clearer path from source documents to a completed return.
See how Soraban fits into your current tax workflow and where it can reduce repeated handoffs without replacing the systems or review steps your team relies on.
Under 30 days to your first live season. No migration. No commitment until you see it working.