One Person, Many Voices: How to Build Agents That Sound Like a Real Person
Most AI writing sounds the same. Confident, generic, generically confident. Feed a model your data and it still comes back sounding like a chatbot pretending to be you.
This session is about what actually closes that gap.
Benjamin Carsley is a founding engineer at The Sentience Company, leading work on personalized voice, tone, and style. His interactive at sentience.com/voices, "One Person Is Many Voices," walks through why the shape of this problem is harder than it looks. On October 16, he'll go into the technical layer underneath: how to build agents that model an individual person's writing, what breaks along the way, and where the current frontier actually is.
In this 60-minute workshop, Ahmet İlten and founding engineers from Sentience, co-present a practical deep dive into building an AI agent that understands its user and communicates like them. Ahmet breaks down the four architectural principles his team learned the hard way, and when each one matters. Ben shows how the agent learns and replicates each user's unique voice, tone, and style.
If you're building an agent right now, this session will give you a clearer framework for deciding what to build, and what not to build.
What you'll walk away with:
Why "sounds like you" is really an evaluation problem, not a model problem
How to structure the voice-and-style layer of a personal agent
What Ben's team is finding in their newest voice-and-style experiments
Who this is for:
Building AI products that need to sound like a specific user
A founder working on personalized voice, writing, or communication tools
An engineer curious about the technical layer underneath personal AI
Format: 60 minutes, live on YouTube. ~40 min technical deep-dive, ~15 min live Q&A.
About the series
Workshop #5 in the TwoSetAI Workshop series.
Session #1 (July 17) Why multi-agent systems fail
Session #2 (July 31) Inside AgensFlow
Session #3 (August 14) Token Budgets for Coding Agents
Session #4 (August 28) Self-Improving Agent Loops
Session #5 (September 18) Building A Personal Agent That Actually Works and Sounds Like You
Session #6 (September 25)
Session #7 (October 2)
Session #8 (October 16)
See all upcoming sessions: https://twosetai.com/lab/workshops/
About the instructors
Benjamin Carsley is an ML and fullstack engineer at Sentience focused on personalized voice, tone, and style. He studied linguistics at Columbia and has worked across data engineering, machine learning pipelines, and speech AI, most recently at BoldVoice. His part of the workshop covers how an agent learns and replicates each user's unique way of writing.
About the host
Angelina Yang runs TwoSetAI and hosts TwoSetAI and Decoding AI Builders, long-form interview platforms where she's interviewed 100+ AI founders shaping this era. fast.ai fellow. 2024 Anthropic Build Tournament winner. Curates the workshop series and connects the speakers with builders who care.
About Sentience
The Sentience Company is a New York-based AI startup building a personal AI model that understands who you are, sounds like you, and operates proactively across the apps you already use. Founded in 2025 by Sam Kececi, the company raised a $6.5M seed round in March 2026 led by Bain Capital Ventures, with participation from South Park Commons, Daybreak Ventures, Terrence Rohan, Soleio, and Annie Case. Sam is a previous guest on TwoSetAI and you'll be able to find his episode on the channel soon.
Hosted by TwoSetAI.
Past events replay
Workshop #1 with Nicole Königstein — Stop Your Agents From Spinning — walked through why multi-agent systems fail in production.
Workshop #3 (Aug 14): Token Budgets for Coding Agents
Workshop #4 (Aug 28): Self-Improving Agent Loops