#29
KPMG shipped 40 fake citations. Anyone checking yours?
Plus: Suralink's full audit AI is live, one control gap remains and SaaS margins are breaking down.
Read this issue →#28
69% plan new billing models. Fewer than 5% have one.
AI cut the research time. The billing model didn't follow. Here's what 1,000 tax pros said.
Read this issue →#27
MassMutual's one rule drove a 30% AI gain
They set a 12-month cap with defined KPIs. If the benchmark wasn't hit, the vendor was out.
Read this issue →#26
The AI gap isn't about tools. It's about tier.
The gap isn't which tools you use. It's whether you redesigned workflows or just layered on top.
Read this issue →#25
One in fourteen finance teams made AI work. Here's how.
Gartner asked 183 CFOs about AI impact. The answer is a 77-point gap that isn't about the tools.
Read this issue →#24
Crypto basis errors: IRS already flagging mismatches
Brokers filed 1099-DAs with zero cost basis for transferred crypto. Here's what CPAs need to do now.
Read this issue →#23
Your firm uses AI. It has no policy for it.
Thomson Reuters surveyed 1,700 accounting pros. Most use AI daily. Most firms have no written rule about any of it.
Read this issue →#22
Your firm's AI usage rate may be measuring noise
KPMG's 75% usage target got gamed. Here's how to build a metric that means something.
Read this issue →#21
42% of finance teams can audit their AI decisions
Three in four finance teams run AI. Fewer than half can explain what it decided or why.
Read this issue →#20
BILL.com cut 700 jobs. Check these 3 things now.
Vendor risk in your AP stack, a Big Four AI signal and the judgment gap no one is talking about.
Read this issue →#19
Copilot reconciles accounts: what to test this week
Two Copilot Finance Agents went live in Excel this week. Here's what to activate before month-end.
Read this issue →#18
Intuit just changed the rules for every CPA firm
ProAdvisor ends in 2027. Check certifications, pricing and client subscriptions before then.
Read this issue →#17
Meta cut teams using AI. Did they prove it first?
Meta cited AI to cut headcount. CFOs who cannot show the same proof are in a weaker position.
Read this issue →#16
The 3 CFO metrics that prove AI is working
Xero's agent builder is here, AICPA audit rules just changed and seat count is the wrong ROI metric.
Read this issue →#15
39% more errors when finance teams run 4+ agents
BCG measured what happens when finance teams run 4+ AI agents at once. The error rate is 39% higher.
Read this issue →#14
Who watches the AI that watches your taxes?
126 systems and no governance plan. The IRS's AI audit expansion has outrun its oversight structure.
Read this issue →#13
PCAOB puts audit AI on the standards clock
The regulator flagged audit tech as a research priority. Standards are coming — document now.
Read this issue →#12
Uber burned its AI budget in 4 months: the CFO fix
Finance teams think seat counts are the win. Uber shows why that breaks before the year ends.
Read this issue →#11
Anthropic built 10 AI agents for finance teams
Claude for Financial Services is the most specific AI-finance spec sheet Anthropic has published.
Read this issue →#10
Finance teams waste 13 hours weekly checking AI
A Sage and IDC study puts a number on how much trust finance pros don't have in their AI tools yet.
Read this issue →#9
The IRS wants to do your tax prep for you
The agency already has your income data. A bill would let filers use it before you see it.
Read this issue →#8
AICPA just made AI literacy a CPA requirement
The AICPA Skills Accelerator is the first formal AI curriculum for CPAs — here's what's in it.
Read this issue →#0
QuickBooks Copilot: 70–90% accurate, one hallucination
We ran it through a real month-end close. Here's what Copilot got right — and where it invented data.
Read this issue →#6
271 bugs found: AI is better at hacking than defense
Anthropic's Mythos found 271 Firefox zero-days — 12x more than Claude Opus. Mozilla is patching fast.
Read this issue →#5
$500 to strip AI safety filters. Cleaning training data doesn't help.
New research shows safety guardrails can be stripped for under $500. Scrubbing training data first doesn't change the outcome.
Read this issue →#4
AI raises your output, lowers your skill ceiling
MIT and Wharton studied this pattern across 30+ papers — and the lag is where the danger lives.
Read this issue →#1
Two AI Labs Restricted Their Best Models This Week
OpenAI and Anthropic both decided their newest AI is too dangerous to release openly. Four stories from a week when capability concerns finally made labs blink.
Read this issue →