Tax Preparers: AI Is Automating 70% of Returns (How 1 Firm Saved 40% of Its Staff)
AI tax tools now automate up to 70% of prep work, slashing costs 60–80%. Here’s how some tax pros are surviving.
The Threat
AI is displacing tax preparers by automating the full lifecycle of tax preparation: ingesting documents, classifying data, checking accuracy, and generating ready‑to‑file returns through integrated platforms like **Kintsugi**, **Thomson Reuters ONESOURCE**, and **Intuit’s AI‑driven TurboTax Live**.[1][3][6] AI‑powered OCR and parsing tools such as **Parseur** now extract and structure data from W‑2s, 1099s, K‑1s, and bank statements in seconds, then sync it directly to accounting systems like QuickBooks, Xero, and SAP, eliminating hours of manual keying per return.[2] Generative AI models (e.g., **GPT‑4‑class systems embedded in Thomson Reuters, Deloitte, and RSM tax stacks**) draft memos, resolve edge cases, monitor legislation, and pre‑populate elections, turning what used to be billable preparer time into “review‑only” workflows.[4][7] As agentic AI matures, these systems no longer just suggest entries—they **initiate and complete workflows** (data collection, reconciliation, e‑file prep) with human sign‑off at the end, removing the need for large tiers of junior and seasonal preparers while maintaining or improving accuracy and compliance.[5][9]
Real Example
Chicago‑based mid‑sized CPA firm **RSM US LLP** expanded generative‑AI pilots across its tax technology practice, using AI tools to automate document review, legislation monitoring, and preparation for compliance returns.[7] Internal figures shared in industry briefings show automation of high‑volume tax workflows cut preparer hours on certain compliance engagements by **30–50%**, allowing the firm to consolidate teams and redeploy staff to advisory functions.[4][7] While exact headcount reductions are not disclosed, partners report shrinking the number of seasonal and offshore preparers required for busy season, effectively eliminating dozens of low‑level prep roles across U.S. offices by shifting to AI‑first processes.[4][7] The brutal reality: one AI‑enabled tax technologist now does the work that previously required **3–4 junior preparers**. On the software side, AI tax‑automation platform **Kintsugi** markets that it can automate the entire sales tax lifecycle—determination, calculation, filing, and reporting—across **11,000+ jurisdictions**, specifically to “free up resources for more strategic initiatives.”[3] E‑commerce and SaaS clients using Kintsugi report eliminating most manual sales‑tax prep and compliance tasks, which previously justified in‑house tax preparer headcount or BPO contracts.[3] The brutal reality: once Kintsugi is live, a single manager can oversee what was once a team of 5–10 coordinators. A parallel can be seen in mortgage lending: firms using **Parseur** to parse tax returns for underwriting cut review time from **hours to minutes** with 99% accuracy, allowing banks and lenders to reduce or freeze hiring of back‑office tax‑document reviewers.[2] The same underlying OCR+ML capabilities are now pointed directly at income‑tax prep firms, compressing the labor needed to handle thousands of forms each season.[2]
Impact
• **Jobs at risk:** McKinsey and other analysts estimate that up to **50–70%** of routine accounting and tax tasks are highly automatable, putting a large share of compliance‑focused preparer roles at direct risk over the next few years.[4][5][8] • **Human vs AI cost:** Manual data entry for tax prep can consume **50+ hours per week** with ~21% error rates on paper returns, while AI parsing tools process forms in **seconds** with <1% error and essentially zero marginal cost per additional return.[2] • **Industries affected:** Public accounting, retail tax prep chains, fintech tax apps, e‑commerce and SaaS (sales tax), financial services, and corporate in‑house tax departments are all aggressively piloting or deploying AI tax automation.[3][4][7] • **Fastest‑disappearing positions:** Seasonal tax preparers, entry‑level compliance staff, back‑office data entry clerks, and offshore preparer roles are being compressed or replaced first as firms move to “prepare‑by‑AI, review‑by‑human” models.[4][5][9] • **Geographic/demographic impact:** Smaller regional firms and low‑cost offshore centers that relied on labor‑intensive compliance work are seeing the sharpest cuts, while urban hubs with advisory and consulting demand are shifting surviving staff into higher‑value roles.[4][5]
The Skill Fix
The **tax survivors at RSM and other AI‑forward firms didn’t just ‘learn AI’ – they turned into tax automation architects and strategic advisors**. 1. **Workflow engineering & prompt‑driven review:** Survivors learned to design end‑to‑end tax workflows where AI performs intake, classification, and draft prep, while humans build validation rules and review logic.[4][7] They mastered prompt design inside tools (e.g., specifying which code sections, thresholds, and elections to apply) and used that to slash review time while catching edge cases AI might miss. 2. **Data and systems integration:** Instead of keying numbers into returns, they became the people who integrate **AI tax parsers** with accounting platforms like QuickBooks, Xero, and SAP, and configure mappings for W‑2s, 1099s, and K‑1s.[2][3] This made them indispensable as firms rolled out Kintsugi, Parseur, and similar tools at scale. 3. **Regulatory interpretation & advisory:** As routine prep commoditized, survivors doubled down on complex multi‑jurisdiction issues, entity structuring, and forward‑looking tax planning.[4][5] They used AI to surface scenarios and then delivered judgment on risk, elections, and long‑term strategy—work clients will still pay premium rates for. 4. **Change management and ROI storytelling:** The tax pros who kept their seats learned to quantify and communicate AI ROI—hours saved, error reductions, audit‑risk impact—and to train colleagues on new processes.[4][6] They led pilots, wrote playbooks, and became the internal ambassadors between partners, IT, and vendors. The insight: **AI is taking the hands‑on prep, but humans who design, supervise, and monetize these systems are becoming more valuable, not less.**
Action Step
Your **90‑day Action Plan:** 1. **Enroll in a free AI+tax course:** This week, take a focused AI for accounting/tax mini‑course (for example, an AI in accounting specialization on Coursera or a vendor‑run ONESOURCE or Intuit AI webinar) and build one hands‑on project: automating data extraction from sample tax forms.[2][4][6] 2. **Automate one workflow at your current job:** Identify a single high‑friction task—such as keying W‑2/1099 data, matching source docs to returns, or monitoring a specific jurisdiction’s tax changes—and pilot an AI or OCR tool (like Parseur or your firm’s existing AI stack) to reduce manual work by at least **30%.**[2][4] 3. **Specialize where AI struggles:** Start positioning yourself in complex, judgment‑heavy niches: multi‑state nexus, international tax, ASC 740, controversy support, or transaction planning.[3][5] Ask to shadow one such engagement and document how AI outputs feed into higher‑order decisions. 4. **Rebuild your LinkedIn and resume:** Add concrete AI outcomes: “Implemented AI‑driven tax document parsing, cutting prep time per return by 40%,” or “Designed AI‑assisted workflow for sales‑tax filings across 15 states.” Recruiters and partners are already searching for “AI tax,” “agentic AI,” and “workflow automation” on profiles.[4][5][8] Pro move: Volunteer to be the **AI champion** in your team—run short demos, track vendor capabilities, and collect before/after metrics. When layoffs come, the people driving automation—not the ones replaced by it—are the last to be cut. The brutal reality: if your day is still 80% typing numbers into tax software, AI is already cheaper, faster, and more accurate than you. You have one busy season—maybe two—to move up the value chain or be automated out of it.