We respect attorney-client confidentiality. No tracking pixels in our emails.
Tax practice has specialized AI needs: IRS ruling analysis, tax memo drafting, transfer pricing documentation, and multi-state research. This guide covers Bloomberg Tax, Checkpoint, and every tool worth adding to your tax workflow.
2026/07/05
When the IRS released Notice 2025-47 in late 2025 — guidance on the tax treatment of certain AI-generated intellectual property assets — tax attorneys at major firms faced a situation that has become routine: a significant new ruling had dropped, clients would ask about it within days, and the team needed to understand the guidance, identify affected clients, and draft preliminary analysis fast. The attorneys who used AI research tools to get a structured preliminary analysis of the Notice in the first few hours after publication arrived at client conversations materially better prepared than those who relied on traditional sequential research.
Tax practice is arguably where the ROI from AI legal research tools is highest and most measurable. Tax analysis is heavily document-based, the source materials are voluminous and well-structured (IRS publications, Treasury regulations, revenue rulings, Tax Court decisions), and the cost of errors — in client advice, filed returns, or formal tax opinions — is concrete and significant. AI tools that can accelerate research, improve memo drafting, and reduce the risk of missing a relevant authority directly reduce the professional liability exposure that defines tax practice risk.
Tax law is among the most research-intensive areas of legal practice, and it has some characteristics that make it particularly well-suited to AI assistance. Tax authorities — the Internal Revenue Code, Treasury regulations, revenue rulings, revenue procedures, private letter rulings, technical advice memoranda, and Tax Court decisions — are voluminous, interconnected, and published in standardized formats. The analytical tasks that tax attorneys perform — determining the tax treatment of a transaction, analyzing whether a structure meets the requirements of a safe harbor, researching the IRS's published position on a specific issue — involve applying a structured body of authority to specific facts.
This structure makes tax research more amenable to AI assistance than areas of law where the relevant authorities are less systematically organized. The AI doesn't need to understand the subjective nuance of a jury's credibility determination or the strategic judgment of litigation posture — it needs to identify the relevant authorities, understand their hierarchy, and apply them to the facts as presented.
The tax research platform market has been dominated by Bloomberg Tax and Thomson Reuters Checkpoint for decades, with each offering comprehensive primary source databases, editorially maintained analysis tools (Bloomberg BNA portfolios, Checkpoint's Federal Tax Coordinator), and practice tools. Both platforms have invested significantly in AI capabilities since 2023, and both released substantively improved AI research and drafting tools in the 2025-2026 period.
The entry of general legal AI tools (Harvey AI in particular) into tax research has created a third option for tax attorneys: using a general-purpose legal AI alongside a traditional tax research platform, getting the depth of the platform's content with the superior synthesis capabilities of Harvey AI's research assistant. This hybrid approach is increasingly common at large law firm tax departments.
Bloomberg Tax has built AI capabilities into its research and analysis platform that go meaningfully beyond search. Its AI research assistant handles tax law questions in a structured way: identifying the relevant Code sections, regulatory provisions, and published guidance, then synthesizing the answer with citations. For standard federal income tax questions — depreciation methods, partnership special allocations, corporate reorganization requirements — Bloomberg Tax's AI research is accurate and well-sourced.
The Bloomberg BNA portfolio system, which provides expert editorial analysis organized by topic, has been enhanced with AI features that allow attorneys to ask follow-up questions within a portfolio context. Rather than searching for the relevant portfolio section and reading it sequentially, attorneys can ask specific questions and receive answers drawn from the portfolio's analysis.
Bloomberg Tax's AI memo drafting tools are further developed for federal income tax questions than for international or state tax. The platform produces solid first drafts of federal tax research memos; international tax memo drafting requires more attorney structuring.
Pricing is subscription-based, typically sold at the firm or department level with per-seat costs that vary based on the specific modules included. International tax coverage, state and local tax coverage, and transfer pricing tools are often separate modules from the core federal income tax subscription.
Thomson Reuters Checkpoint — the incumbent in the CPA firm market particularly — has integrated AI tools primarily through the WG&L library and its federal research workflow. Checkpoint's AI research assistant is well-integrated with the Federal Tax Coordinator (the editorial backbone of Checkpoint's federal tax content) and handles standard federal tax questions effectively.
Checkpoint's AI capabilities are somewhat behind Bloomberg Tax on the drafting and synthesis dimensions in 2026, but the platform's content depth (particularly for CPA-oriented practice tools, state tax research, and international tax) remains a competitive advantage. The integration between Checkpoint and Thomson Reuters' tax compliance software — OneSource — is meaningful for CPA firms that use the full Thomson Reuters stack.
Westlaw Precision integrates Tax Court decisions and some IRS materials, but it is not a substitute for Bloomberg Tax or Checkpoint for serious tax research — the specialized tax editorial apparatus that makes Bloomberg and Checkpoint effective is missing from Westlaw's tax coverage.
Harvey AI's tax research capabilities represent an increasingly viable alternative approach for law firms: use a general-purpose legal AI for synthesis and analysis, with a traditional platform (Bloomberg Tax, Checkpoint, or Westlaw) as the content source for specific authority lookup.
Harvey AI's strength in tax research is synthesizing complex, multi-authority questions. Partnership taxation questions — disguised sales, section 704(c) methods, partnership termination analysis, carried interest — involve interactions between multiple Code sections, Treasury regulations, and a substantial body of IRS guidance and case law. Harvey AI handles these synthesized questions more coherently than the native AI tools in Bloomberg Tax or Checkpoint, which tend to be better at within-topic questions than at cross-topic synthesis.
For international tax, Harvey AI's ability to integrate treaty analysis, foreign tax credit regulations, and Subpart F analysis across multiple authorities is particularly valuable. Tax treaty research is an area where Harvey AI adds meaningful value when used in combination with a treaty database (Bloomberg Tax's international content or a specialized treaty database).
Paxton AI adds value for monitoring — tracking new IRS guidance, revenue rulings, and regulatory developments as they're published. For tax attorneys who need to stay current on IRS releases relevant to their practice areas, Paxton AI's monitoring capabilities translate regulatory publications into actionable practice summaries.
Tax opinions and memos follow predictable structures: facts, issue, applicable authorities, analysis, conclusion. This structure makes them well-suited to AI drafting assistance. The AI can be given the facts and the issue, generate the authority section by identifying relevant Code provisions and guidance, and produce a structured first draft of the analysis.
Draftwise applies to tax memo drafting in its contract and document drafting mode — useful for structuring complex tax opinion language and for maintaining consistency with prior tax opinions. Litera is valuable for the document review and consistency checking functions that large tax opinions require — ensuring that defined terms, Code section references, and cross-references are internally consistent throughout a long opinion document.
For standard tax opinions with predictable structure (qualified opportunity zone opinions, tax-free reorganization opinions, reasonable compensation opinions), AI drafting reduces first-draft time substantially. For novel or complex opinions — those addressing first-impression issues or transactions without clear safe harbor guidance — AI tools assist with structure and authority identification but the substantive analysis requires deeper attorney engagement.
State and local tax research is an area where AI tool coverage is thinner than federal tax research. The volume and variety of state tax authorities — 50 state income tax systems, hundreds of local jurisdictions, constantly changing nexus standards, economic presence rules, and apportionment formulas — creates a research challenge that current AI tools address inconsistently.
Bloomberg Tax's SALT research tools are the most comprehensive available, but even Bloomberg's AI assistance on SALT questions is less reliable than its federal tax capabilities. The error rate on state-specific tax conclusions is higher than on federal questions, and attorneys should apply more rigorous manual verification to AI-assisted SALT conclusions.
Paxton AI monitors state tax legislative and regulatory developments, which is valuable for staying current on nexus rule changes, marketplace facilitator laws, and digital services tax developments across states. This monitoring function is more reliable than asking AI tools to synthesize complex SALT analysis.
Transfer pricing documentation is a specialized area where AI tools are beginning to make inroads but where proprietary tools at the Big Four accounting firms currently have the most sophisticated capabilities. For law firms and smaller accounting firms handling transfer pricing work, Harvey AI's ability to analyze comparable transactions, synthesize arm's length standard analysis, and draft documentation sections is a meaningful productivity tool.
Standard transfer pricing documentation — the Master File, Local File, and Country-by-Country Report required under OECD BEPS Action 13 — follows structured formats that AI can assist with. Harvey AI's ability to analyze client-provided financial data and draft the economic analysis sections of transfer pricing documentation is increasingly used by tax practitioners without access to the Big Four's proprietary tools.
A tax partner needs to quickly analyze a newly released Revenue Ruling affecting partnership distributions and provide preliminary guidance to three clients with relevant positions.
Step 1 — Initial analysis: Harvey AI is given the Revenue Ruling text and asked to: (a) summarize the key holdings, (b) identify which Code sections and regulations are interpreted, (c) flag any departures from prior IRS positions. Output reviewed by partner in 15 minutes.
Step 2 — Client impact assessment: Partner provides Harvey AI with a brief description of each client's relevant fact pattern and asks for preliminary analysis of how the ruling affects each. Output provides starting point for client-specific analysis, which partner then develops.
Step 3 — Authority verification: Bloomberg Tax is used to run citator check on the Revenue Ruling and identify whether any prior rulings it references have been modified, revoked, or superseded.
Step 4 — Client memo drafting: Harvey AI drafts a preliminary client memo structure; partner develops the analysis sections. Final memo is reviewed against Bloomberg Tax's editorial guidance for consistency.
Total time from ruling release to preliminary client memos: approximately 3 hours, compared to a full day under traditional sequential research and drafting.
Harvey AI — Best general legal AI for complex tax research synthesis — partnership taxation, international tax, treaty analysis. Strongest complement to a traditional tax platform.
Paxton AI — Best for monitoring IRS guidance releases and state tax legislative changes. Essential for staying current across a high-volume practice.
Westlaw Precision — Valuable for Tax Court case law research and cross-referencing tax authorities with broader legal context.
Draftwise — Useful for drafting complex tax opinion language and maintaining consistency with prior opinion structures.
Litera — Strong for document review and consistency checking in large tax opinion documents — defined terms, cross-references, Code section citations.
Casetext — Worth evaluating for tax attorneys who need case law research capability integrated with AI analysis at a more accessible price point than Westlaw.
LexisNexis — Relevant for tax attorneys who need Tax Court case law alongside federal regulatory materials within a single platform.
Q: Can AI tools reliably analyze complex partnership tax questions involving multiple interacting Code provisions?
A: Harvey AI handles these questions better than other available tools, but with meaningful limitations. The AI identifies the relevant authorities and provides structural analysis well. For genuinely complex partnership issues — disguised sales with multiple parties, complex section 704(c) elections, partnership mergers — the AI output is a useful starting framework that experienced tax counsel must develop, not a final analysis. The large language model handling of numerical examples and computational tax analysis is less reliable than its doctrinal analysis.
Q: How should we handle attorney-client privilege for tax advice transmitted through AI platforms?
A: Tax advice that would otherwise be privileged as legal advice doesn't lose that protection because it's developed using AI tools. Maintain standard privilege hygiene: ensure AI-assisted work product is developed as part of attorney work product, don't include AI platform vendors in communications that need to remain privileged, and review your platform agreements' data handling terms. The privilege analysis for AI-assisted tax advice is the same as for any computer-assisted legal analysis.
Q: Is there an AI tool specifically for transfer pricing documentation that outside firms can access?
A: Not yet at the level of sophistication that the Big Four have built internally. Harvey AI is the most capable general AI for transfer pricing document drafting currently accessible to outside firms. Bloomberg Tax's transfer pricing tools are the strongest research platform for the economic and regulatory analysis. For high-stakes transfer pricing engagements, the combination of Harvey AI for drafting and Bloomberg Tax for research is the current best available configuration for outside firms.
Q: How do AI tools handle tax due diligence for M&A transactions?
A: AI document review tools — including Harvey AI and some ediscovery platforms — are effective at scanning target company tax documents (returns, audit correspondence, tax sharing agreements, transaction documents) for identified risk indicators. The scope definition work — what to look for — still requires experienced tax attorney judgment. The document review execution — reviewing large volumes of tax records for the identified risk indicators — is where AI tools provide meaningful acceleration.
Q: Are there AI tools that integrate directly with tax compliance software like OneSource or CCH Axcess?
A: Thomson Reuters Checkpoint integrates with OneSource within the TR ecosystem, providing some workflow connection between research and compliance. True bidirectional AI integration between research and compliance platforms is not yet widely available outside proprietary Big Four configurations. This is an area of active product development that will likely see more options in the next 12-18 months.
For research platform selection, the Westlaw vs Casetext comparison is particularly relevant to tax attorneys evaluating primary source coverage.
This article reflects independent editorial analysis. LawyerAI does not accept payment for editorial coverage. Tool scores are based on methodology described in Our 5-Dimension Methodology. Last reviewed: 2026-07-05.