On May 4, Business Insider published an article about KPMG’s new AI dashboard. They’ve been publishing several articles in recent weeks about KPMG’s AI initiatives actually, like the tax simulation tool and this creepy one about how KPMG is preventing Skynet from taking over the firm. The latter was published shortly after the news hit of a cybersecurity group’s AI agent barging into McKinsey’s AI program and, within two hours, gaining access to the full production database including 46.5 million chat messages, 192,000 PDFs, 93,000 Excel spreadsheets, 93,000 PowerPoint decks, and 58,000 Word documents, among other things. You can read about that here if you’re into that sort of thing.
But we’re here to talk about KPMG’s AI dashboard. More specifically, how the firm is requiring AI usage. Let’s start with an excerpt from the Business Insider article:
The firm said it hopes the dashboard, which came online late last year, encourages more “frequent and sophisticated” AI use among the division’s 10,000 workers.
“Our data shows regular AI users produce higher-quality work, feel less stressed, and spend more time on strategic work,” Russ Grote, a spokesperson for the firm, told Business Insider. “These benefits help people progress faster in their careers. They can also better serve our clients who are going through their own transformation programs.”
KPMG says more than 90% of its US employees use AI weekly.
So, about that. Because public accountants do best when you give them actual instructions rather than setting them loose and hoping for the best, the firm is requiring its people to use AI in their day-to-day. Specifically, according to multiple tipsters we’ve spoken to and the BI article itself, they’re expected to have at least one prompt in the firm’s AI tools 75% of working days.
The dashboard tracks some of these tools, as well as external tools like Microsoft 365 Copilot. Many employees are expected to hit a 75% usage target — meaning they’ve used AI on three-quarters of the business days.
One tipster explained to us how they and many of their colleagues are hitting that metric:
Many, including me, would just type random questions into the prompt to hit the once a day quota.
We were also sent a couple posts from Fishbowl where KPMGers react to the rule and explain their own tactics to hit this metric which, as we understand it, is loosely tracked so for all intents and purposes, asking the AI a stupid question or having it go through the motions of summarizing an email you already read is compliance.
Always a good sign when your staff think a rule is a meme.
Funny enough, some staff were already happily using AI and love the stuff but even they find the required metrics silly:
Regardless how you feel about AI and integrating it into your work, I think we can all agree mandatory usage like this is quite silly and a waste of compute.
[/caption]
