Artificial intelligence has moved from conversation to implementation in the tax industry at an unusually rapid pace.
As recently as 2024, it was difficult to identify a single product that could meaningfully assist with preparing a tax return from start to finish. Demonstrations focused largely on summarization, drafting emails, or answering general tax questions. Preparation remained firmly human-driven.
By 2025, that landscape began to shift. Multiple vendors introduced AI-assisted tax preparation tools capable of generating first drafts of Form 1040 returns. The progression did not stop there. In a recent press release, Juno indicated that firms may be able to prepare Form 1065 returns using AI by May 2026. The speed of advancement has been notable.
This rapid evolution raises an important question for firm owners and tax professionals: what can AI realistically do in tax preparation today, and where does it still fall short?
Understanding the distinction is essential. Overestimating AI creates risk. Underestimating it creates inefficiency.
This article examines the practical capabilities of AI in the tax industry today, along with its current limitations.
AI adoption in tax did not begin with preparation. It began with support functions.
Early tools focused on document summarization, data extraction, and research assistance. Over time, machine learning models improved in pattern recognition, particularly when trained on structured tax data. Vendors began building domain-specific AI systems rather than relying on general-purpose language models.
By 2025, several platforms demonstrated AI-generated drafts of simple individual returns. While still dependent on human review, the shift from support to partial preparation marked a turning point.
The development trajectory suggests continued expansion. However, capability today remains uneven across return types and complexity levels.
To evaluate AI’s practical value, it is helpful to separate its strengths from its constraints.
One of the most effective applications of AI in tax preparation is document collection.
The process typically begins with uploading a prior-year tax return into the system. The AI analyzes the return and identifies forms, schedules, and reporting elements previously filed. Based on this analysis, it generates a customized document checklist tailored to that client’s situation.
Instead of sending a generic tax organizer, firms can distribute a dynamic checklist reflecting actual historical data.
The system then sends the checklist to the client through a secure portal. As documents are uploaded, the AI evaluates completeness. If required documents are missing, the system flags the gap and automatically generates follow-up requests.
This reduces manual chasing and improves intake efficiency. The AI does not simply wait for uploads. It monitors progress and prompts action.
In this area, AI performs reliably and consistently. Document intake is structured, repetitive, and data-driven, making it well suited for automation.
Another area where AI has made measurable progress is tax advisory support.
General-purpose language models, including ChatGPT, initially posed challenges in tax research due to hallucinations and outdated information. While they could generate persuasive responses, accuracy was inconsistent. For tax professionals, reliability is non-negotiable.
To address this issue, specialized AI tools were developed specifically for tax research and advisory. Platforms such as Juno Advisory, Blue J, and TaxGPT integrate curated tax databases and authoritative sources into their models.
These systems function conversationally, similar to general AI chat interfaces, but draw from structured tax law references and controlled data sets. As a result, the outputs are more reliable and better aligned with regulatory standards.
In practice, AI advisory tools can assist with:
Human review remains essential, but research efficiency improves significantly.
AI also demonstrates strength in generating structured work papers.
Once data is extracted from client documents, AI systems can organize it into standardized formats. Reconciliations, summary schedules, and supporting documentation outlines can be drafted automatically.
This reduces the time preparers spend formatting and structuring supporting materials. Instead of starting from a blank template, professionals review and refine AI-generated documentation.
Because work paper creation follows predictable formats, AI performs well in this context.
Perhaps the most discussed capability is AI’s ability to draft simple individual returns.
In certain platforms, uploaded documents are analyzed, data is extracted, and a draft Form 1040 is generated. The output is not final. There are often missing elements, classification errors, or incomplete interpretations. However, the baseline draft can reduce preparation time.
It is important to emphasize that this capability currently performs best with straightforward returns. W-2 income, standard deductions, and limited schedules fall within AI’s comfort zone. As complexity increases, performance becomes less consistent.
In all cases, a human remains in the loop. AI produces a draft. A qualified preparer validates, adjusts, and finalizes the return.
The productivity gain lies in acceleration, not autonomy.
Despite progress, AI remains limited in handling medium and complex returns.
Complex returns require contextual judgment. Multi-entity structures, foreign reporting obligations, layered partnership allocations, and evolving business scenarios demand nuanced interpretation. AI systems can struggle with ambiguous documentation or uncommon fact patterns.
As of today, most publicly demonstrated AI preparation tools focus on simple Form 1040 returns. Broader forms such as Form 1065 remain under development.
While vendors project expansion into partnership returns, real-world adoption will depend on reliability. In tax, partial accuracy is not sufficient.
Some AI platforms include review features that attempt to identify inconsistencies or anomalies within returns. These tools can highlight potential errors or missing data points.
However, AI-based review does not yet match the judgment of an experienced tax professional. It can detect structural inconsistencies but may miss strategic issues or nuanced compliance risks.
Review in tax preparation involves interpretation, not only verification. It requires understanding the client’s broader context and anticipating regulatory scrutiny.
AI supports review processes. It does not replace them.
A common misconception is that AI can fully automate the tax filing process.
In practice, AI cannot independently complete the entire cycle. Even if a draft return is generated, a human professional must input or validate data within authorized tax software, obtain required signatures, and initiate electronic filing.
Compliance procedures require identity verification, consent documentation, and professional sign-off. These regulatory safeguards are designed around human accountability.
AI cannot independently assume legal responsibility for filing.
Tax resolution cases require case-specific judgment and negotiation.
When a notice is issued by the IRS or another tax authority, professionals must interpret the issue, determine the appropriate response, draft correspondence, and potentially communicate with regulatory agencies.
This process often involves strategic evaluation and discretionary decision-making. While AI can assist in drafting response letters or summarizing notice details, it cannot independently manage resolution strategy.
Resolution requires human accountability and negotiation.
The consistent theme across both AI capabilities and limitations is the necessity of human involvement.
AI excels at structured, repetitive, data-driven tasks. It struggles with ambiguity, edge cases, and judgment-intensive decisions.
In tax preparation, accuracy carries legal and financial consequences. Therefore, AI should be viewed as an augmentation tool rather than a replacement for qualified professionals.
The optimal model today is human-in-the-loop. AI performs intake, drafting, organizing, and research acceleration. Humans validate, interpret, and finalize.
This hybrid structure balances efficiency with accountability.
AI development in tax is progressing quickly. The shift from zero meaningful preparation tools in 2024 to multiple partial-prep platforms in 2025 illustrates this momentum.
Announcements regarding upcoming capabilities, including potential AI-assisted partnership return preparation, indicate continued investment and innovation.
However, adoption should be grounded in realistic evaluation. Firms should test tools, assess reliability, and integrate AI incrementally. Overreliance on early-stage technology can introduce compliance risk.
Conversely, dismissing AI entirely may result in missed efficiency gains.
The strategic approach lies between skepticism and blind enthusiasm.
AI in tax preparation has advanced from peripheral assistance to meaningful operational support in a short time frame.
Today, AI can:
At the same time, AI cannot:
The future trajectory suggests expanded capability, but present-day implementation requires measured integration.
For tax firms, the opportunity lies in leveraging AI where it performs best while maintaining human oversight where judgment and accountability are essential.
The question is no longer whether AI belongs in tax preparation. It does. The more relevant question is how to use it responsibly within the boundaries of its current strengths.
Understanding what AI can and cannot do today enables firms to adopt it strategically rather than reactively.
How AI can support your tax firm without replacing expertise? Schedule a call to explore smarter ways to integrate AI into your workflow.
When tax firms should consider outsourcing tax preparation to manage workload, improve efficiency, and scale operations during busy seasons.
what organized tax firms do differently during peak season to stay efficient, reduce stress, and deliver better client results.
The real ROI of offshore staffing for CPA firms with our calculator. See how much your firm can save while scaling efficiently.