Founding Infrastructure Program
Get from sign in to a working OrgX operating loop: business context, people, goals, MCP installation, first initiative, decisions, and outcome receipts.
Activation proof
First session should produce work.
< 30 min
First outcome
5 / 10
Slots open
90 days
Operating loop
The founding deployment is measured by visible IWMT progress, not onboarding completion screens.
Day 14 refund guard if the morning brief is not running.
Golden path
Best case: sign in, install MCP, create an initiative, get a receipt.
01
Sign in and define the workspace
Capture business context, goals, owners, operating constraints, and the first initiative that should prove value.
02
Install OrgX where work already happens
Pair the OrgX MCP with ChatGPT, Claude, Cursor, or OpenClaw so agents can read organizational memory and write progress back.
03
Create the first initiative
OrgX turns the ask into workstreams, milestones, tasks, decision gates, and a live execution room.
04
Approve decisions and collect receipts
Humans stay in the loop for budget, risk, and direction. Every accepted output becomes reusable context.
Founder credibility
Built by a technical founder who studies systems that remember.
Hope Atina is building OrgX from the same operating problem the product solves: directed work, durable memory, and human judgment staying connected across tools.
Bioengineering
Trained to reason from living systems: feedback loops, constraints, failure modes, and adaptive structure.
ML thesis
Focused on directed networks, the same shape OrgX uses to keep goals, people, decisions, and work connected.
Operator context
Building OrgX while at Alma, with the founder experience grounded in real operating pressure instead of a detached demo.
MCP install
Bring OrgX memory to the tool you already trust.
The onboarding wizard should detect your preferred surface and guide the right connector path. Chat remains the front door; MCP makes the work portable.
ChatGPT
Add the OrgX MCP connector in Developer Mode, then connect the workspace before starting the first initiative.
ChatGPT can answer from OrgX memory, create work, and surface decision context without losing the thread.
Claude / Claude Code
Install the OrgX MCP endpoint or run the OrgX wizard so Claude can list entities, attach proof, and update IWMT.
Claude sessions can continue work across tasks, PRs, artifacts, and initiative progress.
Cursor / OpenClaw
Pair through the OpenClaw bridge or MCP config, then route code tasks back to OrgX for proof and review.
Engineering work stays connected to business goals, owners, decisions, and launch proof.
Concession-first fit
Keep the best agent tools. Add the operating layer they do not try to be.
OrgX should not pretend to replace the runtime, model workspace, or coding agent. It makes their work durable across Business, People, Goals, decisions, proof, and outcomes.
CrewAI
- Where it wins
- Fast role-based crews, collaborative agent tasks, and workflow experiments when the team wants to prototype multi-agent execution quickly.
- Where it stops
- The run can succeed while ownership, approvals, initiative state, and durable outcome proof still live outside the crew.
- What OrgX adds
- OrgX turns the crew output into initiatives, decisions, owners, artifacts, and receipts that the organization can inspect later.
LangGraph
- Where it wins
- Durable state machines, branching agent workflows, streaming, and human-in-the-loop runtime control for developers.
- Where it stops
- It is a developer orchestration layer, not the business-facing workspace where goals, people, budgets, and operating proof compound.
- What OrgX adds
- OrgX can route work through graph runtimes while keeping the executive surface, IWMT progress, and proof ledger coherent.
Anthropic Skills + Agent SDK
- Where it wins
- Packaging repeatable skills, building Claude-native agents, and extending execution with tools, permissions, sessions, and MCP servers.
- Where it stops
- Skills make the agent more capable, but they do not automatically become shared organizational memory or reusable operating playbooks.
- What OrgX adds
- OrgX promotes repeatable agent work into workspace playbooks connected to tasks, decisions, owners, and verified outcomes.
ChatGPT Projects / Memory
- Where it wins
- Focused conversational workspaces where project memory keeps ChatGPT anchored to the right context and history.
- Where it stops
- Project memory improves the conversation, but it does not by itself assign owners, gate decisions, or prove business outcomes.
- What OrgX adds
- OrgX keeps chat as the front door and writes the work into the organizational graph so context survives beyond the thread.
Claude Projects
- Where it wins
- Long-context research, writing, analysis, and coding work where project files and instructions give Claude a strong local brief.
- Where it stops
- The project can hold context, but completion, review state, accountable owners, and cross-tool receipts still need a system of record.
- What OrgX adds
- OrgX connects Claude work back to the initiative, the people responsible for it, and the evidence that it changed the business.
| Tool | Where it wins | Where it stops | What OrgX adds |
|---|---|---|---|
| CrewAI | Fast role-based crews, collaborative agent tasks, and workflow experiments when the team wants to prototype multi-agent execution quickly. | The run can succeed while ownership, approvals, initiative state, and durable outcome proof still live outside the crew. | OrgX turns the crew output into initiatives, decisions, owners, artifacts, and receipts that the organization can inspect later. |
| LangGraph | Durable state machines, branching agent workflows, streaming, and human-in-the-loop runtime control for developers. | It is a developer orchestration layer, not the business-facing workspace where goals, people, budgets, and operating proof compound. | OrgX can route work through graph runtimes while keeping the executive surface, IWMT progress, and proof ledger coherent. |
| Anthropic Skills + Agent SDK | Packaging repeatable skills, building Claude-native agents, and extending execution with tools, permissions, sessions, and MCP servers. | Skills make the agent more capable, but they do not automatically become shared organizational memory or reusable operating playbooks. | OrgX promotes repeatable agent work into workspace playbooks connected to tasks, decisions, owners, and verified outcomes. |
| ChatGPT Projects / Memory | Focused conversational workspaces where project memory keeps ChatGPT anchored to the right context and history. | Project memory improves the conversation, but it does not by itself assign owners, gate decisions, or prove business outcomes. | OrgX keeps chat as the front door and writes the work into the organizational graph so context survives beyond the thread. |
| Claude Projects | Long-context research, writing, analysis, and coding work where project files and instructions give Claude a strong local brief. | The project can hold context, but completion, review state, accountable owners, and cross-tool receipts still need a system of record. | OrgX connects Claude work back to the initiative, the people responsible for it, and the evidence that it changed the business. |
Business / People / Goals
New users ascend through usage, not tutorials.
Every loop adds context back into the workspace: business facts, responsible people, active goals, decisions, proof, and playbooks.
Day 0
First initiative
One real business outcome, one live execution room, one decision loop.
Week 1
Workspace memory
Business, People, Goals, and operating constraints become agent-readable.
Day 30
Reusable playbooks
The patterns from completed work become templates for the next initiative.
Day 90
Outcome ledger
Your organization has a durable record of what moved, who approved it, and what changed.
Buyer questions
The page should answer fit before a call is booked.
Who is this for?
Technical founders and small leadership teams who already use AI tools, but keep re-explaining business context every time work moves.
What outcome should I see first?
A real initiative with workstreams, tasks, decisions, a live room, and a visible receipt trail from the first meaningful prompt.
How does onboarding work from sign in?
You define the workspace, connect Business, People, and Goals, install MCP where you work, then create the first initiative from chat or command.
Which tools can use OrgX memory?
ChatGPT, Claude, Claude Code, Cursor, OpenClaw, and MCP-compatible clients can read/write through the OrgX MCP surface.
What does OrgX create inside my business?
A reusable operating layer: initiatives, workstreams, milestones, tasks, decisions, artifacts, owners, goals, and proof of outcomes.
How does it connect Business, People, and Goals?
Business context anchors the workspace, People own decisions and tasks, and Goals become the north-star constraints every agent must preserve.
What can block execution?
Missing provider keys, billing credits, integration permissions, unclear owners, and unapproved decisions are surfaced as setup steps rather than silent failure.
What proof shows it is working?
Each initiative should produce visible artifacts, decision receipts, completed tasks, and outcome notes that can be inspected later.
What does founding access cost and how is risk handled?
Founding access is scoped after a proof review. The public managed offer is the Client Proof Loop; if the morning brief is not running by Day 14, the founding cohort refund guard applies.
Founding access
Apply with the context needed to move on day one.
Founding access is scoped after a proof review. The public managed offer is the Client Proof Loop; the first call should start with your business already researched and the first operating loop scoped.