Skip to content

Architecture

Pilots are easy. Production is an architecture test.

Concurrency, longer context and real integrations are what break a voice agent between the pilot and the floor. Voicing answers each of them inside one runtime you can point at.

Book a DemoSee how it runs
  • One runtime, not four vendors
  • Your perimeter
  • Nine layers, scheduled
The runtime01 / 06

One turn, one process, nothing in between.

A stitched stack pays a network hop and a queue at every step of a turn, and the budget is gone before the model answers. Here a whole turn happens inside one process on one machine.

Caller, on your trunkObservability riserTransport and memoryTurn‑taking and V‑VADOrchestration and policyApplication SDKKnowledge and retrievalSpeech captureLLM routerVoice synthesisCPU inferenceConversation store
A section through the Voicing AI runtime: five floors inside one perimeter wall. The call enters from the caller’s own trunk through the roof, runs down the service corridor into the first room of every floor, and the reply returns the same way; every floor writes to one observability riser, and nothing leaves the wall.

  1. Channel

    The trunk ends on your own PBX, so no call leaves your network.

  2. Orchestration

    Policy runs before the model, so consent and redaction cannot be skipped.

  3. Application

    Agents read the systems of record in place, so nothing is copied out.

  4. Model

    Hearing, reasoning and speech are one process, so a turn costs one hop.

  5. Infrastructure

    Ordinary CPU servers, so capacity is racks your team already buys.

Deployment02 / 06

Where the audio sits has an exact deployment boundary.

Every review asks the same question: where does the audio live, and who can reach it. The deployment boundary answers it, and the same image runs on all three.

Deployment02 / 06

Where the audio sits has an exact deployment boundary.

Every review asks the same question: where does the audio live, and who can reach it. The deployment boundary answers it, and the same image runs on all three.

  1. On-premises

    Air-gapped where the review requires it, CPU only, zero egress.

    Telephony and mediaOrchestration and policyAgents and retrievalSpeech and languageCPU computeConversation storeTrunkNo egress port
    Built perimeter. Plan of the runtime standing inside the On-premises perimeter, with six rooms off one corridor and the trunk entering through the perimeter and the building’s own door. Nothing. The trunk terminates on your own PBX and stays there.
    Leaves the wall
    Nothing. The trunk terminates on your own PBX and stays there.
    Stays inside
    Audio, transcripts, weights, traces and the conversation store.

  2. Your cloud

    Inside your own subscription, behind your own IAM boundary.

    Telephony and mediaOrchestration and policyAgents and retrievalSpeech and languageCPU computeConversation storeTrunkYour control plane
    Your account boundary, one port. Plan of the runtime standing inside the Your cloud perimeter, with six rooms off one corridor and the trunk entering through the perimeter and the building’s own door. Your own control plane, billed and logged on your own account.
    Leaves the wall
    Your own control plane, billed and logged on your own account.
    Stays inside
    Audio, transcripts, weights, traces and the conversation store.

  3. Voicing Cloud

    Operated by us, on the same image, in a region you pin.

    Telephony and mediaOrchestration and policyAgents and retrievalSpeech and languageCPU computeConversation storeTrunkOperations, in region
    Named region, one port. Plan of the runtime standing inside the Voicing Cloud perimeter, with six rooms off one corridor and the trunk entering through the perimeter and the building’s own door. Operations traffic, to the region written into the agreement.
    Leaves the wall
    Operations traffic, to the region written into the agreement.
    Stays inside
    Processing, which does not leave the region you named.
Telephony and mediaSIP, RTP, jitter bufferOrchestration and policyConsent, redaction, retriesAgents and retrievalEntitlement filter, citationsSpeech and languageHearing, reasoning, speechCPU computeQuantised weights, no GPU fleetConversation storeAudio, transcripts, tracesOne turn, one processTrunkNo egress portThe runtime, one imageOn-premisesBuilt perimeter, no egress port
Plan of the runtime standing inside the On-premises perimeter, with six rooms off one corridor and the trunk entering through the perimeter and the building’s own door. Nothing. The trunk terminates on your own PBX and stays there.
Telephony and mediaSIP, RTP, jitter bufferOrchestration and policyConsent, redaction, retriesAgents and retrievalEntitlement filter, citationsSpeech and languageHearing, reasoning, speechCPU computeQuantised weights, no GPU fleetConversation storeAudio, transcripts, tracesOne turn, one processTrunkYour control planeThe runtime, one imageYour cloudYour account boundary, one port
Plan of the runtime standing inside the Your cloud perimeter, with six rooms off one corridor and the trunk entering through the perimeter and the building’s own door. Your own control plane, billed and logged on your own account.
Telephony and mediaSIP, RTP, jitter bufferOrchestration and policyConsent, redaction, retriesAgents and retrievalEntitlement filter, citationsSpeech and languageHearing, reasoning, speechCPU computeQuantised weights, no GPU fleetConversation storeAudio, transcripts, tracesOne turn, one processTrunkOperations, in regionThe runtime, one imageVoicing CloudNamed region, one port
Plan of the runtime standing inside the Voicing Cloud perimeter, with six rooms off one corridor and the trunk entering through the perimeter and the building’s own door. Operations traffic, to the region written into the agreement.
Failure modes03 / 06

Four ways an agent fails, each with a control.

The four places an agent programme actually fails, with the consequence of each failure made clear.

  • Decode budgetCaller speechSpeech modelSpoken reply
    Decode budgetCaller speechSpoken replySpeech model

    Speech as spoken

    Trains on disfluent, interrupted and overlapping speech, then answers each turn inside a fixed decode budget.

    A caller who changes their mind is understood.

  • Filter firstEntitled sourcesPermission filterCited answer
    Filter firstEntitled sourcesCited answerPermission filter

    Permissions first

    Retrieval checks this caller’s entitlements before generation, then returns each answer with its source.

    The agent cannot say what this caller may not hear.

  • Schema-strictTool callSchema gateCore record
    Schema-strictSchema gateTool callCore record

    Schema before action

    Each tool call must match the core system’s schema before it is sent. Failed arguments are retried or handed on.

    A wrong field never reaches your core system.

  • Riser, all floorsFloor tracesTrace riserTrace store
    Riser, all floorsFloor tracesTrace riserTrace store

    One trace, every floor

    Turns, tools, retrievals, handoffs and model versions reach a store in real time, so one trace holds each call.

    A bad call is traced to its own cause, not guessed at.

Ownership04 / 06

A model changes. Your call flow should not.

Each layer carries two lines, with Voicing first and a stitched stack second. Read them together and the difference is what can change without you.

A stitched stack is the usual alternative: a hosted voice API for the calls, a speech vendor for the transcript, a frontier model provider for the reasoning and a second synthesis vendor for the reply.

Solid, with Voicing.

Dashed, with a stitched stack.

One runtime.

Nine layers end here, on infrastructure you control.

  1. 01Telephony ingress

    With Voicing: YouEnds on your own PBX

    With a stitched stack: The vendorA rented number, not yours

  2. 02Speech recognition

    With Voicing: VoicingTuned models, on your disk

    With a stitched stack: The vendorEvery second is metered

  3. 03The language model

    With Voicing: VoicingSelf-hosted, pinned by you

    With a stitched stack: The vendorShared, no pin, no budget

  4. 04Voice synthesis

    With Voicing: VoicingOur voices, same machine

    With a stitched stack: The vendorSecond hop, second invoice

  5. 05Orchestration and tools

    With Voicing: SharedYour policy, our runtime

    With a stitched stack: SharedTwo vendors and your glue

  6. 06Knowledge and retrieval

    With Voicing: YouIndexed inside the wall

    With a stitched stack: SharedCopied to an outside index

  7. 07Conversation data

    With Voicing: YouKept on your retention

    With a stitched stack: The vendorKept on their retention

  8. 08Evaluation and traces

    With Voicing: YouTraced into your own store

    With a stitched stack: The vendorA sampled view per vendor

  9. 09Release cadence

    With Voicing: SharedYou approve the window

    With a stitched stack: The vendorChanged on their schedule

Your perimeter

At scale05 / 06

Built for the floor at peak, not for the demo.

Concurrency, context and integration are where an agent programme loses its value. These are the figures a capacity review asks for.

What scale costs an agent programme

Up to 58%

The fall in Agent Value Multiple as a programme moves from pilot to full scale and adds concurrency, complexity and context, per Gartner.

  • 700 ms

    Measured across BFSI, on‑premises, real telephony, 80th percentile

  • 97%

    Function calls on 2,800 calls, schema‑strict grading, eight‑tool suite

  • Under 0.5%

    Hallucination rate on in‑scope intents, measured in production

  • $0.017

    Per minute on CPU, on‑premises, at 10 m minutes per month

The capacity model, CPU footprint and concurrency curve come to the architecture review on request.

Review06 / 06

Bring your architect. Bring the hard questions.

A ring of three segments, blue, sage and sand, on a stand inside a drawn perimeter; a cloud and a server rack wait at docks outside it.
Architecture

The runtime, on your infrastructure

A working session with an engineer who has deployed inside a bank’s perimeter. We map your telephony, data boundary and failure modes on this same topology, and tell you what we would not automate.

Book a Demo