19 articles
Mercor gave 12 experienced CPAs every fact they needed, in the files, and they still scored 37%. The machine found all of it, every time, and 34 times faster even on the items the humans got right. That isn't a test of your accountants. It's a test of whether what a client already told your firm is anywhere your process actually looks, and in most firms it isn't.
Xero's AI has coded 100 million transactions, and practitioners trust exactly one of its four methods: the one built from written rules. Every coding engine on the market infers your client's intent from history. Not one of them will read it, even though you could write it down today.
Anthropic just put a month-end close agent inside your client's QuickBooks. The close packet it produces is designed to land on your desk. The question is whether your engagement is priced for what the agent replaced — or for what it can't do.
Your first agent shouldn't live inside your accounting platform. Start where the door is already open.
2026 is the year of the agent — and you're five minutes away from building one you could reuse every day. An agent isn't a new technology. It's the same AI you're already using with a better job description. The gap between a prompt, a workflow, and an agent is smaller than most CAS practitioners think.
Most of the tools CAS practices pay for aren't doing anything AI can't do — they're formatting engines that take data you already have and present it the way someone expects it. Two real engagements proved the point: a QBO year-end that normally required a ledger in Xero, and a working trial balance request that normally required Caseware. Both replaced in an afternoon with purpose-built agents.
The KPIs that built successful accounting firms — utilization, AHR, profit per partner — were designed for a world where humans do all the work. When AI becomes a production participant, those metrics aren't wrong. They're just no longer enough. Here are three metrics I think matter now — and I want you to challenge me on every one of them.
Anthropic surveyed 80,508 AI users. The number one fear isn't job loss — it's unreliability. AI fails the same way your staff does. The difference: corrections become permanent rules.
Intuit and Anthropic announced a multi-year partnership putting QuickBooks inside Claude. The MCP server is already public on GitHub — a developer expanded it to 143 tools. But the depth question that's limited every platform integration still applies, and the Enterprise Suite demo isn't what your QBO clients will get.
Marketing teams are seeing 90% cost reductions. Customer success teams are doubling productivity. And Anthropic's own data shows financial roles have an 85% theoretical AI capability with only 20% actual adoption. The gap between other industries and CAS practices isn't closing. It's widening — and it's measurable.
Anthropic measured what AI actually does across 800 occupations — not what it could theoretically do. Financial roles show 85% theoretical capability but only 20% actual adoption. That gap isn't reassurance. It's a closing window, and the firms moving now are the ones who'll own the other side of it.
Sequoia Capital named accounting a $50–80B disruption target. Their framework shows where CAS practices are exposed — and what's still defensible.
