The IRS just put AI use in tax practice on the record

Banks client data directly into your AI assistant

Lili's MCP server connects Claude and other compatible AI tools directly to client Lili bank accounts, letting you query cash positions, flag unpaid invoices, and spot estimated tax shortfalls across your whole book in a single prompt — no portal-switching. Why it matters: If you're already running a morning Claude workflow, this replaces the manual data-pull step and gets you to the actual advisory question faster. Catch: MCP integration requires a compatible AI assistant and client accounts already on Lili — if your clients use Chase, BofA, or anything else, this does nothing for you today. Unknowns: Lili's 30%+ time savings claim has no independent validation, and pricing for the MCP integration isn't broken out publicly — ask before piloting. My take: Niche play until MCP connectivity expands to major banks, but worth 30 minutes if your clients are already on Lili. What it signals is more interesting than the product itself: bank data piped directly into AI workflows is coming, and Lili is just first to ship it. [Read more →]

Managing your cap table shouldn’t drain your time or derail your budget. Pulley takes the complexity and surprises out of equity management with intuitive workflows, audit-ready compliance, and accurate reporting—all backed by support and pricing you can rely on.

CFOs and finance leaders trust Pulley to stay compliant with key tasks like 409A valuations, ASC 718 reporting, or 83(b) elections without incurring expensive legal fees or getting bogged down with manual work. With Pulley, you get 100% audit-defensible data, powerful tools, and real expertise.

Learn more here

Pilot Claims Full Month-End Close, No Humans — Pilot launched Meridian, an AI platform built on 187,000+ months of bookkeeping across 8,000 businesses that it says executes the entire month-end close and returns review-ready financials — pitching one accountant managing a $1M book as the new unit economics. Why it matters: Firms evaluating outsourced close should put Meridian on the shortlist, but press Pilot hard on what "review-ready" means in practice before committing — it's a claim, not a spec. (QuickBooks Online only right now — firms on other ledgers get limited functionality.) [Read more →]

Kinter Launches AI Agents Inside Your ERP — Kinter, backed by a16z, Bain, and YC, publicly launched AI agents that autonomously prepare accruals, identify prepaid expenses, and draft journal entries directly inside NetSuite and QuickBooks, with early customers reporting up to 70% time savings on expense-side close work. Why it matters: For firms with clients on NetSuite paying for manual close work, this is worth a demo call — but the 70% figure is self-reported, and automation covers only the expense side of close for now. [Read more →]

40% of US Workers Used AI to Fake Expense Receipts — An Emburse survey of 2,000 US and UK workers found 40% of US employees have used AI to generate fraudulent expense receipts, with 19% fabricating purchases outright — and 40% of those used their employer's own AI tools to do it. Why it matters: The receipt-image review your clients call an internal control is no longer sufficient — if you're doing advisory work on expense policy, that conversation now has to include AI-detection tools. (Emburse is a spend-management vendor with direct commercial interest in the findings.) [Read more →]

Weekly AI Use for Tax Research Doubled in a Year — Blue J and CPA.com's survey of 1,000+ US tax professionals found weekly AI use for tax research jumped from 33% to 60% in 12 months, with 90% still relying on general-purpose tools like ChatGPT rather than tax-specific platforms. Why it matters: ChatGPT doesn't cite primary sources — if a client or examiner asks how you reached a position, "I asked ChatGPT" isn't a defensible answer, and the IRS disclosure item below is the direct regulatory consequence of exactly this behavior. (Blue J is a tax-specialized AI vendor that co-authored the survey.) [Read more →]

You don't have to scroll every AI thread, track every new tool, or watch every demo. 

The Rundown AI breaks it all down for you — the latest AI news, tools, and tutorials in one free 5-minute email every morning. 

Trusted by 2M+ professionals at Apple, Google, and NASA.

Sign up to start learning.

If the IRS's new AI alert prompted you to check your client data handling, start with your prompt history — anything you've typed into a non-enterprise AI tool containing a client name, EIN, or dollar figure is a potential data exposure. Most ChatGPT free and Plus accounts train on conversations by default unless you've turned it off in settings. [Read more →]

⭐ IRS Moves Toward Mandatory AI Disclosure for Return Preparers

The IRS Advisory Council recommended mandatory AI disclosure requirements for return preparers, state legislatures are introducing bills targeting AI in professional services, and AICPA chair Jan Lewis publicly stated AI will take over basic tax prep — all in the same week. These aren't independent signals; they're the beginning of a coordinated regulatory response to the 88% adoption figure that Intuit published this week (see above). Small firms that build a documented AI governance framework now — what tools you use, how you review outputs, how you bill — get to design it themselves. Firms that wait will get a compliance form to fill out instead. Why it matters: If you're using AI in any part of return prep and you can't describe your review process in writing, a disciplinary complaint would expose that gap immediately. [Read more →]

Deloitte Puts 85,000 Auditors on Agentic AI Platform — Deloitte launched a unified agentic intelligence network inside its Omnia audit platform, connecting AI agents across risk identification, evidence analysis, documentation drafting, and compliance review at global scale. Why it matters: When Big Four firms standardize agentic audit delivery at this scale, mid-market and small-firm audit pricing comes under margin pressure — firms without an efficiency story will face that conversation from audit clients within 18 months. [Read more →]

Only 21% of CFOs Rate Their AI Readiness as Advanced — EY — EY's study of 1,600+ finance leaders found 21% rate their AI readiness as leading or advanced, with 47% saying their teams can't measure value from emerging technology investments. Why it matters: If your advisory clients can't measure AI ROI, that's a billable gap — and this EY benchmark gives you a credible third-party frame for the conversation. (EY sells AI readiness consulting, so the framing of a widespread readiness gap reflects their commercial interest.) [Read more →]

The convergence of IRS guidance, mandatory disclosure proposals, and adoption surveys hitting 88% in a single week makes one thing clear: the informal AI experiment phase at most small firms is over. The ones that document their review process now do it on their own terms.

—Alex

Speak your prompts into ChatGPT or Claude and get detailed, paste-ready input that actually gives you useful output. Wispr Flow captures what you'd cut when typing. Free on Mac, Windows, and iPhone.

Try Wispr Flow free

📬 Know an accountant drowning in manual bookkeeping? Forward this.