01
Service and product concierges
Help someone explore an offering, ask follow-up questions, and find a useful next step. Ground the conversation in approved information and make the limits of the service clear.

Project Sapiens builds custom voice agents and conversational interfaces for businesses. We connect spoken requests to useful information and actions, combining conversation design, OpenAI technology, and hands-on product development.
Tell us what you want to buildIndependent expertise. Hands-on delivery. Member of the OpenAI Partner Network.
What we can build
For teams building a voice-enabled product, a service concierge, or a conversational way to get work done. We help define where voice is useful, what the experience needs to handle, and when a person should take over.
01
Help someone explore an offering, ask follow-up questions, and find a useful next step. Ground the conversation in approved information and make the limits of the service clear.
02
Let a user ask for information or prepare an action through conversation. Confirm important details before updating a record or passing a request to another system.
03
Add a conversational interface to a web or mobile product. Design how people start, interrupt, correct, and leave the conversation alongside the rest of the experience.
From prototype to production
User journeys, representative dialogues, clarification behavior, and clear paths for errors and human handoff.
Voice integration and agreed connections to business information and tools. OpenAI’s Realtime API is one foundation we use for responsive conversation.
Evaluate interruptions, background noise, names, corrections, response time, and task completion using representative scenarios.
An agreed deployment approach, documentation, and a plan for monitoring quality, cost, and the conversations that need review.
The people behind the product
Founded by Tony Marks, Project Sapiens brings his experience in conversational AI, voice interfaces, and digital products to the work of building useful software. Earlier Sapien AI projects reflect Tony’s work and the work of that team. Engagements can take the form of focused projects, fractional AI leadership, or ongoing advisory.
Yes, where suitable sources and access are available. We scope retrieval and integrations around approved information, then test how the agent handles missing or conflicting answers.
Phone-based experiences can be scoped alongside web and app experiences. We assess telephony requirements, provider compatibility, consent, and handoff before agreeing on the implementation.
We design for clarification, repetition, correction, and confirmation. Testing includes realistic interruptions and unclear requests, with a route to a person or another channel when needed.
Yes. A pilot can test one useful conversation and the integrations behind it. We agree on scope, success criteria, timing, and fees before work begins.
Also exploring Custom AI agents? We can scope the experience and the work behind it together.
New projects & collaborations
Tell us what you’re trying to build or improve, who it’s for, and where you are today. We’ll review your inquiry and get back to you.
An early idea is welcome. Budget and timing can be approximate.
Prefer email? tony@projectsapiens.xyz