Skip to content
PERRY DOLMANS AI BUILDER & CREATIVE TECHNOLOGIST
WORKLABNOTESABOUT Get in touch

← BACK TO WORK

AI WORKFLOW · PRODUCT

Private AI Venture Studio

From rough idea to structured product plan.

STATUS
BUILDING
ROLE
Concept, architecture & build
TOOLS
Claude Code · specialised agents · GitHub
YEAR
2026
Private AI Venture Studio

A private AI operating system that turns rough ideas into validated, structured product concepts.

01

The idea

Most ideas die in the gap between a thought and a plan. I wanted a private studio of specialised AI agents that could research, validate and structure an idea the way a small team would — but on demand and under my control.

02

The problem

Generic chatbots forget context and skip the boring, decisive steps: market checks, scope, architecture, phasing. I needed something that remembers the project and works through it in order.

03

The approach

Discrete agents with clear roles — research, validation, architecture, planning, quality — coordinated through structured documents and project memory, so each step builds on the last instead of starting over.

04

How AI was used

AI does the legwork; I make the calls. It drafts research, proposes structure and flags risks, and I decide what is real.

Product discovery and market research
Idea validation against real constraints
Architecture and phased planning
Quality control on its own output
05

What I learned

Specialised, role-based agents with shared memory beat one clever prompt. Structure is what makes AI output trustworthy.

CURRENT STATUS

Actively building. Used on my own projects, including several featured here.

NEXT STEPS

Tighter validation gates and reusable templates so a new idea reaches a structured plan even faster.

Interested in how this was built?

Get in touch See more work