Discussion about this post

User's avatar
Hisashi Miyaguni's avatar

The doubled ROI number is the one I would want pulled apart, because it is the kind of finding that changes what a firm does on Monday.

Firms that write a policy are firms that were already organised enough to write one. So the comparison may not be policy versus no policy. It may be a firm that writes things down versus a firm that does not, and writing things down is also what lets you report ROI at all.

Which raises a harder possibility. A firm with no policy may have the same return and simply be unable to see it, because it never wrote down what it was trying to get. If that is what is happening, the finding is measuring visibility rather than returns, and the advice that follows is still right, write the thing down, but for a different reason than the number suggests.

I am one of your nine in ten. We run outsourced accounting operations for 29 clients in Japan and we have no documented AI policy. We published our four control questions and answered no, no, no and partially. That is what the absence looks like from the inside, and it took publishing it to notice.

Does the report separate firms that measured ROI from firms that reported it?

No posts

Ready for more?