Walk through the platform—from natural-language research and client workspaces to automated review and tax-document OCR.
Natural language, purpose-trained on the tax code. Every answer anchored in primary authority — ready to paste into a client memo or audit defense letter.
Can Meridian defer gain on the sale of their Denver warehouse if they use a qualified intermediary but won't identify the replacement property for 60 days?
Yes. Under a Section 1031 exchange, Meridian can defer gain on the Denver warehouse sale using a qualified intermediary (QI), even if the replacement property isn't identified immediately.
The key timing rules under IRC §1031(a)(3) 1 are:
At 60 days, Meridian would exceed the 45-day identification window, which would disqualify the exchange under Treas. Reg. §1.1031(k)-1(b) 3.
Ask the way you think — multi-step reasoning, follow-ups, and clarifications, just like talking to a senior partner.
Every response surfaces the exact IRC section, regulation, ruling, or case it relies on. Click to verify the source.
Federal, state, and municipal in one query. Compare treatment across jurisdictions without switching platforms.
Type / to pull up purpose-built commands — draft a memo, email a client, pull a deadline, or kick off a full return review.
Draft a client-ready research memo from the current conversation
Compose a client email summarizing findings and next steps
Look up the correct IRS or state form for a specific filing
Pull filing deadlines and extension dates for an entity
Side-by-side comparison of tax treatment across jurisdictions
Initiate a multi-stage automated return review
Organize entities, related parties, and engagements — with full data isolation between clients and your firm-wide policies. Answers that understand your client's facts, not just the tax code.
Manage your firm's client entities and engagements
Scope the conversation to a single entity. Tax position, prior returns, supporting documents — all in the context window, all the time. The AI knows your client before you ask the question.
Drop in returns, statements, K-1s, and workpapers. The OCR pipeline parses every page and indexes the facts — so you can ask questions about what's in them.
Consolidated view · TY 2025
Bank, P&L, balance sheet, K-1s, prior returns — unified into one live snapshot. Spot the story in the numbers before your client calls.
Kick off a full review with one click. Multi-stage analysis runs in the background — missed deductions, open issues, and risk flags surfaced before filing. Every finding traces back to the source and the authority.
2025 Form 1065 · Partnership Return
Equipment purchase of $127,400 on 3/15 may qualify for immediate expensing under §179.
Partner K. Chen received $45,000 not reflected on Schedule K-1, Line 4c.
Asset class 39 using SL/39yr but placed-in-service date suggests bonus depreciation eligibility.
Partnership income exceeds threshold; W-2 wage limitation under §199A(b)(2) not computed.
Colorado income apportionment uses single-factor sales; verify sourcing methodology.
Data validation, income analysis, deduction review, compliance check, and risk assessment — each runs independently and reports findings.
Each finding is classified by severity with the specific line item, dollar amount, and authority cited. High-impact items surface first.
A clean completion summary you can hand to a client or reviewer. Every finding traces back to the source and the authority.
TaxScan is tuned for tax documents. It extracts fields, tables, layouts, multi-page forms, and handwriting so document facts can be used in downstream TaxFigure workflows.
TaxFigure TaxScan vs. Industry Leader · IRS Form Suite
| Benchmark category | TaxFigure | Comparison baseline |
|---|---|---|
| Field Extraction | 100% | 73.1% |
| Table Parsing | 100% | 61.2% |
| Layout Accuracy | 99.8% | 68.5% |
| Multi-Page Forms | 99.9% | 52.3% |
| Handwriting OCR | 97.2% | 44.8% |
| Output Usability | 100% | 58.7% |
Internal OCR benchmark using an IRS form suite. Percentages are category scores; higher is better. Results may vary by scan quality and form type.
Control how the AI behaves across your firm
Automatically detect and tokenize PII (SSNs, EINs, account numbers) before sending to AI models. Client data never leaves your vault.
PrivacyVault, firm-wide guardrails, per-role access. Your client data stays yours — and stays out of anyone else's training set.
By the Numbers
Decision guide
The right choice depends on the work, the level of authority required, and whether research needs to stay connected to client documents and review workflows.
| Approach | Best for | Strengths | Typical tradeoffs |
|---|---|---|---|
| TaxFigure | Tax teams that want cited research, client documents, and review workflows in one system. | Tax-specific queries, source links, entity context, document processing, and firm controls. | Requires professional verification and delivers the most value when client work is organized in entity workspaces. |
| Traditional tax research database | Practitioners who need an established tax library and editorial analysis. | Deep research collections, familiar authority hierarchy, and editorial material. | Research may remain separate from client documents and downstream review work; training and licensing vary by provider. |
| General-purpose AI assistant | Early brainstorming, plain-language explanations, and non-client-specific drafting. | Flexible prompts and low-friction drafting across many topics. | Tax authority, freshness, citations, and sensitive-data handling must be checked for the specific service and response. |
| Manual primary-source research | Narrow or high-stakes questions that require direct review of controlling authority. | Direct control over source selection, interpretation, and documentation. | Requires tax expertise and can involve searching across multiple agencies, jurisdictions, and document formats. |
Individuals can begin with a research question. Firms can add teammates, create client entity workspaces, set access, and upload the documents needed for entity-scoped work.
TaxFigure supports research and review; it does not replace professional judgment. Verify authority, dates, extracted facts, and client-specific conclusions before filing or advising a client.
Use primary-source research for final verification. An established research library may fit teams centered on editorial material. General AI may fit low-risk, non-client-specific drafting.
Common Questions
The mechanics behind research, documents, insights, and security.
Ask in natural language — a quick lookup, an eligibility question, an entity-structure comparison. Answers lead with a bottom-line-up-front summary, carry inline citations, and include a sources panel plus suggested follow-ups.
Files move through a three-stage pipeline — queued, processing, processed. OCR extraction and field classification run automatically, then the document becomes available to chat, insights, and the entity's document checklist. PDF, DOCX, XLSX, TXT, CSV, JPG, and PNG up to 10 MB.
Anchor a conversation to a specific client — individual, LLC, S-corp, partnership, or trust — and answers reflect that entity's structure, tax year, and uploaded documents instead of generic guidance.
Insights produces an entity overview with an A–F health grade, a document coverage matrix, and per-document analysis modules (bank statements, P&L, balance sheets, prior-year returns, K-1s). Deduction Hunter scores deduction opportunities by confidence and dollar value.
Yes — slash commands produce structured outputs like research memos and citation lookups, and conversations persist with history and pinning so engagements pick up where they left off.
PII is detected and tokenized before reaching a model, entity workspaces are isolated per client, team members get configurable per-entity access, and accounts support multi-factor authentication.
Answers current as of 2026-11-05.
Join tax practitioners who've stopped losing hours to manual research.