Portal
Twin Factory Portal
Thesis, twin gallery, scorer, ROI, architecture, intake — plus the live Control Plane (/twin-portal/control-plane) for inbox, review, and runs.
This is the working surface for the SolidCage thesis — a hub of live demos, ROI tooling, readiness diagnostics, and the decks behind them. Pick anything below; it all loads in seconds.
Build your digital twin · Open the Twin Factory Portal · Score an agent ROI
Four steps, under 20 minutes total, that take you from curious to running a governed agent.
12 questions across 5 dimensions. Find out exactly where your org sits before touching a single tool.
An AI interview that produces a deployable agent-spec.yaml — identity, guardrails, and a skills bundle — ready for the factory.
Run a use case through the ROI generator. Get an opportunity score, top risks, and a 30/60/90 roadmap you can bring to the table.
See the agentic factory in motion — inbox routing, HITL review gates, and multi-twin runs all in one view.
Every link below is a working demo running on this same deployment.
Portal
Thesis, twin gallery, scorer, ROI, architecture, intake — plus the live Control Plane (/twin-portal/control-plane) for inbox, review, and runs.
Twin Builder
10-minute AI interview that generates a deployable digital twin package — agent-spec.yaml, identity, guardrails, and skills — ready for agent-factory.
ROI
Score an agent use case and get a 30/60/90 roadmap.
Scorecard
12-question / 5-dimension organisational readiness check.
OI Tracker
Track the rate of improvement on the metrics that matter.
Three decks summarising the playbook, the investor case for AI employees, and the engineering pattern behind digital twin factories.
Playbook deck
How to operate an organisation as an agentic OS.
ROI deck
The investor-grade case for governed AI employees.
Twin Factories deck
The engineering pattern for compiling a role description into a production agent.
Mirror docs and agent-readable specs for each layer of the stack.
github.com/fszale/agentic-playbook
The PROMPT layer — context, rules of engagement, operating playbooks.
github.com/fszale/agent-kernel
The AGENT layer — execution loop, tools, memory, planning.
github.com/fszale/digital-twin-filip
The TWIN layer — a digital twin modeled on the founder.
github.com/fszale/agent-factory
The FACTORY layer — multiple twins routed by role with HITL gates.
Filip Szalewicz helps engineering and operations leaders stand up a governed agentic OS — from first digital twin to a routed agent factory with HITL review gates.
Real results from the playbook
Numbers are real, names are anonymized — drawn from the Agentic OS Playbook case studies (slides 11–13).
Burn rate reduction
SaaS startup · 4 humans + agent factory
Feature delivery time
SaaS startup · post-acquisition
AI employees to humans
E-commerce agentic OS · HITL + guardrails
Curated skills deployed
200-person engineering org