What happens when an AI answers a veterinary clinic's phone?

Do callers hang up on an AI receptionist?

During business hours, 1.06% of calls ended with the caller hanging up without saying anything. On the remaining calls, the caller spoke. The speaking measure asks a direct question: did the caller say anything at all during the call?

For a practice considering an assistant on its phone line, this is a useful place to begin. Before a request can be understood or passed to the team, the caller has to contribute to the conversation. In the observed business-hours calls, that happened on 98.9% of calls.

The distinction is between a call with caller speech and a call that ends without it. Keeping those groups separate gives the practice a clear view of what happens at the start of the interaction. It also identifies the silent-hang-up group for the follow-up analysis below.

On business-hours calls, 98.9% of callers spoke and 1.06% hung up without a word.

Business-hours speaking behavior, March–September 2026.

Does it hold up month to month?

Yes. The business-hours speaking rate stayed between 97% and 99% in every month from March through September. As the panel took on new clinics, each month's share remained within that range. The high overall speaking share was accompanied by a high share in every month observed.

The after-hours result showed a similar pattern: 1.35% of callers hung up without a word, compared with 1.06% during business hours. Silent hang-ups were uncommon in both groups. The finding therefore extends beyond the business-hours total to the calls reaching the after-hours assistant.

Consistency is the useful point here. The monthly results show that speaking remained common as the observed panel grew; the business-hours and after-hours comparison shows a low silent share in both settings. Together, they give a fuller account of how calls began than the overall speaking percentage alone.

What happens after a silent hang-up?

In September, 12% of text-backs sent after silent hang-ups received a reply. The rate uses messages sent as its denominator. It describes a written response following a call on which the caller had said nothing.

For those responding, the contact continued in writing. The phone record ended without caller speech; the follow-up message gave the recipient another opportunity to respond. A reply marks the start of that written exchange.

Speaking and replying belong to different stages of contact. The speaking rate describes engagement during an eligible assistant-answered call. The text-back rate describes a response within the separate group sent a follow-up after a silent hang-up. A practice reviewing these results should keep those stages visible in its records.

How we measured this

The speaking measure records whether the caller said anything during an eligible call; a silent hang-up contains no caller speech. Both analyses use the eligible clinic panel and count client calls only: clinic-side lines and calls ended by a line fault are excluded. The monthly range describes each month's business-hours call share from March through September as the panel changed, rather than following a fixed group of clients. The September text-back rate is replies as a share of text-backs sent after silent hang-ups, using a separate eligible follow-up subset.

The unit is a call or a follow-up message, not a unique person. Someone who called repeatedly could contribute more than once. The displayed speaking and silent-hang-up percentages retain their source precision and may not add to an exact total because of rounding. See full definitions and exclusions in the methodology.

What this can't tell you

Speaking is engagement, not a satisfaction score. There was no satisfaction survey, and these results do not establish how many callers would have preferred a person. They also do not establish why a caller hung up silently, whether AI performed better than a human-only alternative, or whether a text reply resolved the original request. The panel describes clinics using PupPilot's assistant.

Measure it at your clinic

Calculate calls on which the caller spoke ÷ eligible assistant-answered calls × 100. Record silent hang-ups separately from disconnects after speech begins. Apply documented exclusions consistently, including the clinic-side-line and line-fault exclusions used for this panel.

Repeat the speaking calculation by month and separate business-hours and after-hours assistant groups. Retain the eligible-call count behind each percentage and note when clinics, routing, or inclusion rules change.

For text-back, calculate follow-up messages receiving a reply ÷ eligible text-backs sent after silent hang-ups × 100. Keep the sent-message denominator, reporting period, and reply window explicit, and link each reply to the message that prompted it.

Report the speaking and text-back rates side by side as separate measures. That gives the team a repeatable account of where conversation begins: during the call or through the follow-up sent after a silent hang-up.