This is a straightforward case study about economics. A professional services firm was spending 14 staff-hours per week on three categories of admin work. We deployed three agents. The work is now done automatically. Here are the numbers.
The 14 hours
The time broke down as follows: 6 hours per week on client onboarding paperwork (collecting information, creating accounts, sending welcome packs), 5 hours per week on invoice follow-up (checking payment status, sending reminders, escalating overdue invoices), and 3 hours per week on meeting scheduling (back-and-forth coordination across time zones).
Agent 1: onboarding (6 hours → 25 minutes human review)
The onboarding agent receives a new client confirmation, sends the information collection form, processes the responses, creates accounts in the CRM and project management tool, and sends the welcome pack. A human reviews the completed pack before it goes out — 25 minutes of review instead of 6 hours of processing.
Agent 2: invoice follow-up (5 hours → 0 human time)
The invoice agent monitors payment status in the accounting system, sends reminders at 7, 14, and 28 days overdue, and escalates invoices beyond 35 days to a human for relationship management. The escalation is the only human touchpoint — and it is only for the cases that genuinely need human judgment.
Agent 3: meeting scheduling (3 hours → automated)
The scheduling agent handles the calendar coordination: receiving scheduling requests, checking availability, proposing times, handling rescheduling, and sending confirmations. It handles 90% of scheduling without human involvement.
The economics
The three agents cost $647 per month. The 14 hours of staff time, at the firm's fully-loaded cost, was costing approximately $2,100 per month. The net saving is roughly $1,450 per month — a payback period of less than two months on the implementation cost. The staff time freed up was redirected to billable client work.