Abby measures its published plans in Abby Minutes: one minute of human answering uses one Abby Minute, while one minute of AI answering uses half an Abby Minute. Estimate your expected mix before comparing the plan allowance with another answering service.
Independent guide from Sysevo, a competing provider. We are not affiliated with Abby. This is not a hands-on product review. Official sources checked on 1 November 2026; verify current terms directly.
Published plan amounts and allowances
The official Abby pricing page checked for this guide shows the following monthly plans. Confirm the current offer and contract directly. The allowance is a branded accounting unit, so do not interpret it as the same amount of human and AI talk time.
| Plan | Published monthly amount | Abby Minutes |
|---|---|---|
| Starter | $164 | 50 |
| Essential | $329 | 100 |
| Professional | $599 | 200 |
| Growth | $1,380 | 500 |
Model the human and AI portions separately
Using the published unit definition, an illustrative month with 100 human minutes and 200 AI minutes consumes 200 Abby Minutes: 100 plus half of 200. The total conversation time is 300 minutes, but the allowance consumed is 200 units.
For a mixed call, ask how the transition and each portion appear in the usage records. The pricing source describes human and AI portions billed at their respective rates. Confirm timing, rounding and any applicable exclusions for the exact arrangement rather than treating a mixed call as entirely one category.
Decide which calls need which answering path
Build a list of routine questions, new enquiries, existing-customer administration and requests requiring staff authority. Define which can begin with AI, which need a receptionist and which should reach your team. Ask to see the actual configuration for those routes.
Do not set the mix solely to minimise billed units. If a caller repeatedly reaches the wrong path, staff may spend more time fixing the handoff. The useful design pairs each call category with an outcome and an accountable owner, then checks the resulting customer experience.
Test what happens at the handoff
The AI Receptionist product page describes AI and human handoff options. Confirm the route included in your proposed service. A transfer to your own employee and a handoff to the provider’s receptionist create different staffing responsibilities.
Use fictional enquiries to test both a normal request and one that needs a person. Ask the receiving person to continue without the builder explaining the situation. Inspect the contact record, summary and follow-up task as well as the connected call.
Ask how the allowance behaves in a busy month
Provide the expected human/AI mix and a higher-volume version of the same workload. Request the actual overage calculation and any plan-change terms in writing. A rate derived by dividing a subscription by its allowance is not necessarily the rate used for extra units.
Include retained phone costs, implementation and staff work outside the service. Use the same time period, tax basis and currency when comparing another proposal. Keep unknown costs visible until resolved instead of entering zero by default.
Measure completed requests and corrections
During a pilot, separate calls answered from bookings completed, messages correctly assigned and issues resolved. Track cases that return to staff and the time spent correcting them. Changing the AI/human mix may alter both the bill and the work remaining with your team.
For an alternative model, review Sysevo voice agents, CRM and current plans. Bring the category list to a demonstration and request a proposal with the same required outcomes and responsibilities.
Questions buyers ask
Are Abby Minutes identical to conversation minutes?
They depend on the answering mix. The published definition counts human time at one unit per minute and AI time at half a unit per minute.
How should a mixed answering proposal be compared?
Estimate the human and AI portions separately, apply the confirmed billing rules and include the staff work remaining after each handoff.
What should be tested before changing the answering mix?
Test which requests reach each path, the context at handoff and the saved next action. Check the effect on customer completion and staff correction time.
Related buying guides
