customer_support · workflow
Building voice AI at scale: Assembled's multi-agent architecture for millions of production calls
Running voice AI reliably in production is far harder than building demos: achieving high conversational quality, low latency, and near-100% uptime simultaneously — while working with rapidly evolving technology — creates significant engineering challenges.
How it works
Common implementation structure
How this type of workflow is generally built, generalized across documented cases — not tied to any one vendor's stack. Click any stage to read what happens there. Specific products that implement these stages appear in “Tools commonly seen” below.
Stage 1 · Customer initiates call
A customer calls in and the voice agent begins handling the interaction.
Tools used
GPT-4o miniGemini 2.5 Flashspeech to texttext to speech
Outcome
Assembled's voice agents handle millions of calls per year in production using a multi-agent architecture, with deep quality auditing enabling continuous improvement in conversation quality and reliability.
What failed first
A single-prompt approach to voice agents — putting all context into one LLM prompt — failed to scale to enterprise production volume, and real-world deployment exposed problems with poor audio and frustrated callers that could not be anticipated from demos.
Results
Volumemillions of calls per year
Grounding & classification
Source type: technical build writeup
24 fields verified against source quotes.
agentic workflowconversational aimulti agent workflowspeech to textvoice aicall recordingknowledge basefailure mode describedhuman review describedproduction runtime claimedtools describedworkflow describedsoftwarethroughput increasetechnical build writeupcall center aicustomer supportagentic task executionescalation workflowvoice call handling