Insurer Accounts Payable: Cycle Time and Automation Sequencing
A regional insurer's accounts payable unit receives 2,200 invoices monthly. An intake clerk sorts and scans them (average 1.5 minutes each). Three AP specialists code and key invoices (average 6 minutes each), but 22% are returned for a missing PO or cost-center coding and require a 4-minute rework loop. A supervisor approves invoices over $5,000 (18% of volume, 2 minutes each). Invoices wait in batches; average total cycle time is 9.4 working days, while average touch time is often quoted as 'about 25 minutes.' The CFO wants straight-through processing for low-risk invoices and a 50% cycle-time reduction in two quarters. IT can automate data extraction but not coding decisions. Staff resist because last year's workflow tool was abandoned after go-live. The unit has no baseline for first-pass yield.
1. Calculate the average touch time per invoice and the process cycle efficiency. Then identify the largest controllable delay and explain why the '25 minutes' figure is misleading.
Read the worked answer
Touch time = intake 1.5 + coding 6 + rework (0.22 × 4 = 0.88) + approval (0.18 × 2 = 0.36) = 8.74 minutes per invoice. The quoted 25 minutes overstates touch time by about 2.9×; it likely includes waiting or an older process, so it should not be used to set improvement targets.
Cycle efficiency = touch time / cycle time. Assuming an 8-hour working day, 9.4 days = 9.4 × 480 = 4,512 minutes. Efficiency = 8.74 / 4,512 = 0.19%. Waiting consumes over 99% of the elapsed time. The largest controllable delay is batching at intake and supervisor approval queues, not the 6-minute coding step. Coding is the largest touch element, but it is small relative to wait. First-pass yield is 78% because 22% require rework. Confirm the working-day assumption and measure queue time by stage before committing to the 50% target.
2. Recommend a redesign and automation sequence. Address the 50% cycle-time target, the IT constraint, and staff resistance.
Read the worked answer
Do not automate the current process. First eliminate batching: scan on receipt into a digital queue, pre-fill vendor and coding data from the PO at intake, and route missing-PO exceptions to a separate queue. Use IT extraction for header data and three-way matching; route only exceptions to specialists, since IT cannot make coding decisions. Define straight-through processing rules for low-risk invoices: PO present, known vendor, amount below a threshold, and no prior exception.
The 50% cycle-time target is plausible mainly by cutting wait, not touch. Pilot one invoice category for 30 days, measure first-pass yield and queue time, then expand at 60/90-day gates. Address resistance by involving specialists in rule design, publishing exception reasons, keeping a rollback path, and measuring quality rather than headcount. Two quarters is tight; if the pilot does not reduce wait by at least half, renegotiate the target rather than automate a broken flow.