Treat AI agents like employees: give them human names, rate their performance on a 1-to-4 scale, and track token costs to make sure they're cheaper than hiring a person. Monitor ROI the way you would for any staff member.
I am a big believer that agents must be named like employees, so humans. And so I would pick whoever first name at tabs.com is still available. I don't think we yet have a Regina, so let's go Regina at tabs.com. I put them on a 1 to 4 scale, which is a top performer. So I'd probably put them at a 2. I don't think they're a 1 anymore. Last year, 1 is typically someone who's on the lower end and struggling, but I think Regina is like probably a 2. But I want to make sure Regina is not costing me more than an employee. So I'd also track their token utilization to make sure that the ROI in Rowena is good enough so that it's not only better than employee, but I'm also getting some real ROI out of it.
Trust and Leadership Lessons for FP&A From 100+ Top Leaders with CFO Jon Laudie
