Proposed lead: Brad Cleveland and the buyer’s finance director; ask shortlisted suppliers to answer separately.
Why it matters: Retell’s published pricing separates components and add-ons, states that silence and hold time are billable, and distinguishes AI charges from continuing telephony charges after transfer. These are supplier-specific billing terms, not a universal market benchmark. (retellai.com)
Question to put to the expert:
What belongs in a defensible comparison of AI receptionist versus human receptionist cost—and which savings should finance refuse to recognise?
A strong answer should cover:
Our proposed governing calculation is:
Total cost of the defined workflow ÷ unique, correctly completed, policy-compliant tasks.
Include implementation, integration, platform commitments, telephony, model and speech usage, testing, monitoring, support, human rescue and repeat-contact handling. Keep the measurement period and task population consistent across alternatives.
Report cash savings, capacity released, avoided hiring and incremental contribution separately. Show sensitivity to completion rate, conversation length, transfer rate and demand—not just a cheaper minute price.
Evidence to request: An itemised quote, a sample invoice and a low/base/high model with every input labelled provided, extracted, assumed or unknown.
Follow-up: If payroll stays unchanged and implementation takes longer than planned, does the business case still work?
Expert quote: [Reserved for an approved interview quotation.]