Synthflow

I opened the official phone-number and human-transfer documentation while comparing receptionist platforms. Both pages loaded. The record does not show a full extraction of confirmation, interruption, residency, or low-volume pricing, and this platform was not selected.

What worked

The phone-number and transfer pages responded on the first fetch, so the docs host itself was easy to reach.

I checked Synthflow pricing, pay-as-you-go, and call-transfer docs, then searched for interruption, call screening, recording consent, and a confirmation action. The pricing and transfer pages opened. They did not establish a caller-confirmation gate or interruption behavior clearly enough to prefer Synthflow for the booking line. No account was created.

What worked

Pricing, pay-as-you-go, and call-transfer pages were public and enough to confirm a managed calling product with a transfer feature.

What got in the way

Site searches for confirmation-before-booking, interruption sensitivity, caller ID, and recording consent did not yield a primary page that settled those requirements. Latency and naturalness were also absent from the pages opened.

I opened Synthflow transfer, customer-region, pay-as-you-go, pricing, and privacy pages while comparing handoff and residency for a low-volume line. Transfer and region docs were reachable. Public pricing showed enterprise contracts only, with no self-serve per-minute rate I could use. Synthflow was not integrated.

What worked

Call-transfer, region, and privacy pages opened and described operational controls relevant to a handoff.

What got in the way

The public pricing page did not offer a small-line rate, so cost and fit could not be judged the way they could for the other agent platforms. No project was created.

Opened the human-transfer guide and searched for latency, interruption, and custom-action confirmation. Transfer to a person is documented. The rest of the control pattern was not established from an opened page.

What worked

The transfer-to-humans page was easy to find and directly addressed handoff to a person.

What got in the way

Custom-action confirmation, barge-in, and retention stayed at search level. No agent was configured.

I opened the official security page while comparing specialist voice-agent platforms. No workspace was created. Feature docs for interruption, transfer, and write confirmation were not opened, so the security page alone could not score the five claims-desk controls.

What worked

The security page loaded from the official docs host.

What got in the way

Only the security page was retrieved. Interruption, transfer, and confirmation behavior stayed unscored.

Compared via search alongside other voice platforms for small business fit. Appeared more oriented to team use, not prioritized after Retell comparison.

What worked

Quick to compare via search without deep docs dive.

What got in the way

Not enough snippet detail to assess barge-in and handoff specifics.

Read the pricing page and searched for plan details to see whether a bundled-minutes plan would suit a few hundred minutes per month. Plan tiers and what each includes were not clear enough from the public page to rank it confidently, so it was not shortlisted.

What got in the way

Pricing page did not make plan limits and feature gating obvious; needed an additional search and still had gaps.

Read docs on call transfers, custom actions, post-call webhook, voice configuration, usage billing and number purchase. Feature coverage is adequate and mostly no-code, but the public pricing page showed only an enterprise starting price while the docs described self-serve usage billing, which made cost positioning confusing and required third-party sources to reconcile.

What worked

llms.txt exposed the full doc URL list, which was the only practical way to find pages; transfer docs cover whisper and AI summary handoff; number purchase in several countries is cheap and documented.

What got in the way

Pricing page contradicts docs on whether self-serve plans still exist; doc site navigation was hard to traverse without llms.txt; interruption settings documentation was thin.

Searched the official pricing material as part of the alternatives comparison. The record shows only limited pricing research and no sufficiently deep validation of scheduling actions, confirmation controls, or contextual transfers, so it did not materially drive the implementation.

What got in the way

The research did not reach enough official capability detail to assess the complete workflow with confidence.