AI Agents & Automation
-
AI Evals for Accountants: How to Know If Your Output Is Good Enough to Use
Most CPAs are approving AI output with a skim. Here’s a structured five-step method — an AI eval — to actually know whether your work product is good enough to send.
-
AI Clutter & Cognitive Load at Work
A peer-reviewed study of 34 finance professionals found that AI clutter hurts output quality 3x more than task complexity. Here’s what accountants need to know.
-
AI Agents Need Good Data First
Why data infrastructure — not AI tools — determines your AI ROI. Learn the three pillars every enterprise needs before deploying agents.
-
AI’s Task Horizon Problem: What the Latest Research Tells Accountants
METR’s January 2026 update confirms AI capability doubles every ~3 months — but a 50% success rate on 5-hour tasks is far from enough for accounting and finance workflows. Here’s what the data means for you.
-
AI Gets a C+ in Accounting: What the 2026 Benchmark Really Means for CPAs
DualEntry tested 19 AI models on 101 real accounting tasks. The best scored 77.3%. Here’s what that means for how you use AI in your practice.
-
Agentic AI in Accounting Firms: Real-World Case Studies
AI adoption in accounting firms jumped from 9% to 41% in one year. Here’s what firms like RSM, PKF O’Connor Davies, and CLA are actually deploying — and the results they’re seeing.
-
AI Output Etiquette for CPAs: How to Stop Sending Sloppypasta
AI tools have made writing free — and that’s created a new professional problem. Here’s what CPAs need to know about ‘sloppypasta’ and how to use AI responsibly.
-
AI’s Last Mile Problem: Why Pilots Succeed But Transformations Fail
Most large enterprises are pilot-rich but transformation-poor. Here’s why AI productivity gains don’t show up on the balance sheet — and what it means for accountants.
-
Skills Are the New Software
Learn what Agent Skill files are and how a single markdown file can make AI work the way your firm actually works — reliably, repeatedly, every session.
-
The 8 Levels of Agentic Engineering (And Where Finance Teams Fit In)
A new framework maps out 8 levels of AI use — from tab complete to autonomous agents. Here’s what it means for accountants and finance teams.