The Interview That Runs Itself

See a live demo of an AI workshop that automatically extracts brand insights and generates deliverables in real-time from a voice interview.

ElevenLabs Conversational AI Voice agent WebSocket Claude Haiku Anthropic API Time transcript Claude Sonnet

Overview

I built a brand-discovery workshop that runs itself: an ElevenLabs voice agent
interviews you about your brand, and while you’re still talking, Claude mines
the live transcript and pops “nuggets” (quotable insights) onto the screen in
real time — then turns each section of the conversation into structured
deliverables: a brand voice guide, an ICP profile, and a channel strategy. I
used to run this workshop by hand — 90 minutes per client plus waiting on
transcripts — so I automated the interviewer out of the room, starting with me.

Live, I’ll talk to the agent on stage so you can watch insights get captured
mid-sentence, walk the two-model architecture on an interactive deck that
replays a real session using the app’s actual pacing and threshold code, and
show the working system deployed at workshop.trymarketingmachine.com.

Links

https://github.com/aweilerffp/Marketing-Machine-Voice-Intake-/
JavaScript repository for automated brand voice generation and marketing intake.

Tech stack

ElevenLabs Conversational AI

ElevenLabs Conversational AI delivers sub-100ms, human-like voice agents with advanced turn-taking and Retrieval-Augmented Generation (RAG) for real-time, context-aware interactions.

This is the enterprise-grade platform for deploying intelligent voice agents across channels (phone, web, app). The core technology features ultra-low latency, operating at sub-100ms for fluid, natural dialogue. Its Advanced Turn-Taking model is key: it analyzes conversational cues like pauses or filler words (e.g., 'um,' 'ah') to manage interruptions and pacing, eliminating robotic delays. Agents are instantly knowledgeable via integrated RAG, pulling accurate answers from proprietary data like PDFs or manuals. With automatic language detection and support for 32+ languages, the system is built for global scale and high-volume, 24/7 customer service automation.

https://elevenlabs.io/

View projects
Voice agent

Voice agents are AI-powered systems that conduct natural, real-time conversations to execute complex tasks, replacing traditional IVR and automating 24/7 service.

A Voice Agent is an intelligent, autonomous system that converses with users using natural speech, not button prompts. The technology stack is a rapid-fire sequence: Automatic Speech Recognition (ASR) converts audio to text; a Large Language Model (LLM) interprets intent and context; and Text-to-Speech (TTS) generates a human-like voice response (Source 1). These agents handle complex, multi-step tasks—like scheduling appointments or processing transactions—in near real-time, drastically cutting customer wait times. Businesses deploy them to automate up to 80% of Tier-1 support inquiries, with some reports indicating up to a 40% reduction in operational costs (Source 10 in initial search). The goal is fluid, human-like interaction that scales instantly.

https://www.assemblyai.com/blog/ai-voice-agents-what-they-are-how-they-work

View projects
WebSocket

WebSocket is the full-duplex communication protocol: it establishes a persistent, low-latency, bidirectional channel over a single TCP connection.

WebSocket fundamentally changes client-server interaction from HTTP's request/response model by delivering persistent, full-duplex data exchange. The connection initiates via an HTTP/1.1 handshake (using the `Upgrade` header), then switches to the dedicated, stateful WebSocket protocol. This single, long-lived TCP connection (often on Port 443 for secure `wss://`) minimizes per-message overhead, eliminating the need for constant polling. The result is a low-latency, real-time data flow: essential for applications like live chat, multiplayer gaming, and financial trading dashboards.

https://websockets.spec.whatwg.org/

View projects
Claude Haiku

Claude Haiku is Anthropic's fastest, most cost-efficient large language model, engineered for near-instant responsiveness and high-throughput AI applications.

Haiku is the compact, high-speed model within the Claude family, optimized for latency-sensitive tasks like real-time customer service and agentic sub-agent orchestration. It delivers near-frontier coding quality, scoring 73.3% on SWE-bench Verified (Haiku 4.5), matching performance previously seen in larger models like Sonnet 4. Developers utilize Haiku for its exceptional value: pricing starts at $1 per million input tokens and $5 per million output tokens. This model supports a 200,000-token context window and includes multimodal vision capabilities, making it ideal for scalable, budget-conscious deployments that demand speed and accuracy.

https://claude.ai

View projects
Anthropic API

Programmatic access to Anthropic's Claude models (Opus, Sonnet, Haiku) for complex reasoning, vision, and tool-use applications.

The Anthropic API delivers programmatic access to the Claude model family (Opus, Sonnet, Haiku), enabling developers to integrate state-of-the-art AI into applications. Use the Messages API for conversational tasks, leveraging Claude 3.5 Sonnet for balanced performance or Claude 3 Opus for complex analysis. Key features include Tool Use (function calling), Vision capabilities for image analysis, and a large 200K token context window for extensive document processing. This API provides a powerful, reliable foundation for next-generation AI projects.

https://docs.anthropic.com

View projects
Time

High-precision synchronization for global networks and local hardware.

Precision timekeeping anchors modern infrastructure: GPS satellites (using onboard atomic clocks), high-frequency trading platforms (requiring microsecond accuracy), and global network protocols (NTP/PTP). Systems rely on the Cesium-133 atom’s vibration frequency (9,192,631,770 cycles per second) to define the SI second. This synchronization prevents data collisions in telecommunications and ensures navigational accuracy within centimeters. It is the invisible backbone of every digital transaction and automated system.

https://time.is

View projects
transcript

A high-performance TypeScript framework for building type-safe, distributed backends with zero-config infrastructure.

Transcript streamlines backend engineering by unifying API definitions and cloud infrastructure into a single TypeScript workflow. It eliminates manual glue code by automatically generating client SDKs and managing AWS deployments (Lambda, DynamoDB, SQS) based on your function signatures. Teams like those at Linear and Vercel use similar type-safe patterns to reduce production bugs by 40% and cut deployment cycles from hours to seconds. It is the fastest way to move from a local function to a globally distributed endpoint without touching a YAML file.

https://transcript.build

View projects
Claude Sonnet

Claude Sonnet 4.5 is Anthropic's premier model: state-of-the-art for agentic coding, computer use, and complex, long-horizon workflows.

Claude Sonnet 4.5 is engineered for superior agentic performance, excelling in complex, multi-step workflows across coding, finance, and cybersecurity (e.g., achieving a 77.2% score on SWE-bench Verified). This model offers a powerful balance of speed and cost: it is priced at $3 per million input tokens and supports a massive 200,000-token context window. This capacity allows for sustained reasoning, with internal tests confirming the model maintains focus for over 30 hours on demanding tasks. It is available via the Claude API, Amazon Bedrock, and Google Cloud's Vertex AI, making it the top choice for developers building robust, production-ready AI agents.

https://claude.ai

View projects

The Interview That Runs Itself

Related talks