The short answer

Budget separately for setup, subscriptions, usage, maintenance and human review. Estimate low, typical and busy-month volumes. Time saved is a capacity estimate; it becomes cash savings only when an actual expense changes.

Separate the cost types

Request separate amounts for process mapping, implementation, integrations, testing, training and support. For recurring charges, identify the unit: seat, task, conversation, minute or another metered action. Ask whether retries, test runs and extra environments count. Record the currency, tax treatment and billing period so the offers are comparable.

Measure the current work

Count how often the task happens and how much active staff time each instance takes. Include corrections and chasing. Then estimate how much reviewing and exception handling remains after automation. Do not count an entire employee salary as a saving when only a small part of their week changes.

Work through an example

Consider an illustrative workflow with 400 enquiries per month. At 6 minutes each, manual handling takes 40 hours. If 320 enquiries need 1 minute of review and 80 exceptions still need 6 minutes, the new workload is 800 minutes, or 13 hours 20 minutes. The estimated capacity released is 26 hours 40 minutes per month. These are sample assumptions, not client results.

Compare cost with capacity

Using an illustrative value of US$20 per staff hour, that released capacity is worth about US$533 per month. Suppose software and maintenance total US$150 monthly and setup costs US$1,200. The estimated monthly capacity value after running costs is about US$383, giving an indicative payback of 3.1 months. This is not a cash-saving forecast: if your payroll remains unchanged, the benefit is time available for other work. Taxes, transition effort and opportunity costs are not included.

Test the busy month

Recalculate with higher volume, longer calls, more failures and more human review. Ask how spend is capped or alerted and what the system does when limits are reached. If the workflow is only worthwhile under optimistic assumptions, reduce the scope or postpone it.

Make the proposal testable

Agree which outcome a pilot must demonstrate and how it will be measured. Include a process for handling failures, a support owner and an exit plan. Compare the pilot with your measured baseline before extending the commitment.

Your next step

Send your workflow, monthly volume, current tools and a provider quote if you have one. We can help identify the assumptions and costs to check.

Explore the related solution

Keep reading