VioBuilding your agentAnalysis, actions & fulfilment

Analysis, actions & fulfilment

View as Markdown

These three tabs decide what happens to a Vio call’s outcome: Analysis scores sentiment and whether the call met its goal, Actions can hand the live call to a human, and Fulfilment POSTs the finished call to your own systems. Together they turn a phone conversation into structured data you can act on.

Analysis — score every call

Turn on Analyze every call and Vio reads each finished call for two things:

  • Sentiment — how the caller felt overall.
  • Goal met — if you provide a success criteria rubric (a short description of what a successful call looks like), Vio judges whether the call met its goal, returning goal met or not met plus a short reason.

Results appear on the individual call in Call logs, so you can filter your day’s calls by sentiment or by whether they hit their goal.

Actions — transfer to a human

On the Actions tab, toggle on Transfer to a human and the agent can hand a live call to a person. Set:

  • A target phone number in E.164 format (e.g. +919876543210).
  • An optional “when to transfer” instruction telling the agent the conditions to hand off — for example when the caller asks for a human or the topic is out of scope.
⚠️

Transfer requires your carrier to support call transfer. If your number can’t transfer, the agent can instead schedule a callback as a fallback.

Fulfilment — results go to your webhook

This moved. The per-agent “Results webhook URL” that used to live on this tab has been replaced by workspace-level webhooks, set up once under Settings → Webhooks. The Fulfilment tab now just points you there.

Register an HTTPS endpoint once and it receives events from every agent you own — each event says which agent it came from, so one endpoint covers your whole workspace instead of a URL pasted into each agent.

Subscribe to call.completed and you get, the moment a call ends: the call metadata, the AI summary, the extracted variables, and the analysis (sentiment, goal-met) — everything you need to drop the call into a CRM, a sheet, or an automation in Zapier / Make / n8n.

What the workspace-level version adds over the old per-agent field:

  • A signed request you can verify, so you know the payload is genuinely from us.
  • Automatic retries with backoff when your server is briefly down.
  • A delivery log you can inspect, rather than guessing whether an event was sent.
  • A verification ping at registration, so a typo’d hostname fails loudly there and then instead of silently swallowing a month of events.

Set up webhooks for the payload shape, the signature check, and worked verification code in Node, Python and PHP.

If you configured a per-agent results webhook before this changed, it is no longer in use — re-register the URL under Settings → Webhooks to start receiving events again.

How the three fit together

TabTurns onOutput lands in
AnalysisSentiment + goal-met scoringCall logs and the webhook analysis{}
ActionsLive transfer to a humanThe call itself (during the call)
FulfilmentNothing per-agent — set up once in Settings → WebhooksYour CRM / sheet / automation, via call.completed

Frequently asked

Q. What do I put in the success criteria rubric? A one- or two-line description of what a successful call looks like for this agent — for example “Caller shared budget and agreed to a site visit.” Vio uses it to judge goal met / not met and writes a short reason.

Q. When does the call.completed webhook fire? When the call ends. Vio POSTs a single call.completed event containing the metadata, AI summary, extracted variables, and analysis — to every endpoint in your workspace subscribed to it. Register one under Settings → Webhooks; see Webhooks.

Q. Does the agent need anything special to transfer a call? Your carrier must support call transfer. Set the target number in E.164 format, and optionally tell the agent when to hand off.

Q. Where else can I see sentiment and goal-met without a webhook? On each call in Call logs. The webhook is only needed to push the data into your own systems.

Next

Tune per-agent call behaviour in Advanced settings, or review completed calls in Call logs.