The AI Operating System Maturity Model: From Scattered Tools to a Firm That Runs on One System

1. Why a Model, Not Just a Score

A score tells you where you are. A maturity model tells you where you are and what to do next, which is the only reason to measure.

This is a model for the firm itself, not its portfolio companies. For scoring the companies you own, use the portfolio AI maturity assessment. The question here is narrower and more useful for the GP: how far along is our own firm in turning AI from a pile of tools into a system we run on. Four stages, and most firms are earlier than they think.

2. What This Measures

This measures whether your firm has an operating system or a pile of tools, and how far between the two it sits.

It is not how many AI subscriptions you hold. A high number is often a symptom of stage one, not progress. It is whether AI is connected, standardized, governed, and genuinely how work gets done. A firm with one well-run, connected workflow is more mature than a firm with fifteen disconnected tools, and the model is built to make that obvious.

Maturity here is about integration and permanence, the same line from what is an AI operating system, turned into a ladder you can place yourself on.

3. The Four Stages

Four stages, from a pile of tools to a firm that runs on a system.

The stages are cumulative. You do not connect live data (stage three) before you have shared Projects and a data rule (stage two), the way you do not automate a process you have not standardized.

4. Stage 1: Scattered

You are at stage one if your firm's AI lives in individual subscriptions and personal habits. One person likes ChatGPT, another uses Claude, a third bought a niche tool. Nothing is shared, nothing is standard, and nobody can tell you what the firm spends or whether deal data is sitting on consumer accounts.

Stage one is risky even though AI is doing something. The value walks out the door when a person leaves, and the confidential-data exposure is real and unmanaged. The way out of stage one runs through a commercial plan, a data rule, and shared Projects, which is stage two. More tools only deepen the pile.

5. Stage 2: Standardized

You are at stage two when the firm runs on a commercial plan, the data rule is set, and shared Projects hold the firm's context so everyone works the same way. This is where most of the easy value lives, and a firm can sit here productively for a while.

The limit of stage two is that the AI still only knows what you paste or load. It does not see your live systems, and nothing runs on its own. The move to stage three is connection: wiring the assistant into your real data through MCP and turning a workflow into something that runs.

6. Stage 3: Connected

You are at stage three when AI reaches your real systems (the data room, the CRM, the portfolio data) and at least one important workflow genuinely runs on it, not as a person's manual chore. Screening happens against live criteria. Monitoring reads live numbers. The system has started.

Stage three is where the firm pulls ahead, because the work compounds: each connection and workflow makes the next one cheaper. The risk is stalling here with one connected workflow and calling it done. The move to stage four is breadth and depth: more of the firm's core work running on the connected system, governed and reliable.

7. Stage 4: Embedded

You are at stage four when the firm runs on the system. The core of how you source, diligence, monitor, and report flows through connected, governed, largely automated AI, deployed where you control it. People do the judgment. The system does the assembly, continuously.

Almost no firm starts here, and not every firm needs to reach it. But this is where AI stops being a productivity tool and becomes a structural advantage, the difference a buyer or an LP can see. It is what the AI Operating System is built to deliver.

8. The Five Things That Move You Up

Five things determine which stage you are in. Score each honestly, on evidence.

Plan and governance. Consumer accounts, or a commercial plan with rules and controls. Context. Do shared Projects hold the firm's knowledge, or does everyone start blank. Connection. Is AI wired to your live systems, or only to what gets pasted. Workflows. Do core workflows run on it, or is it ad hoc help. Ownership. Is there a named owner and a plan, or is it everyone and no one.

A firm is only as mature as its weakest dimension. Brilliant Projects on personal accounts with no governance is still stage one, because the exposure caps you there.

9. How to Place Your Firm Honestly

Place your firm by its weakest link, not its best demo.

The common self-deception is to point at the one impressive thing (a partner's clever Claude setup) and call the firm advanced. Ask instead: is deal data on consumer accounts anywhere (stage one), can the AI see your live systems (below stage three), does anything run without a person (below stage four). The honest answer is usually a stage lower than the instinct.

Far from discouraging, that gap is the most valuable thing the model gives you, because the distance between where you are and where you could be is exactly your opportunity, named.

10. The Move From Each Stage

Each stage has one move, and naming it stops the model from being a chart that changes nothing.

From scattered: get on a commercial plan, set the data rule, build shared Projects. From standardized: connect one live system and make one workflow run. From connected: extend the connected, governed system across more of the firm's core work. From embedded: keep the lead and turn it into the exit or LP story.

The move follows from the stage. You do not need to leap to the end. You need to take the next step well, which is the discipline of the rollout playbook.

11. Where to Start

Place your firm this week, honestly, on the four stages. It takes an hour and it is almost always a stage lower than the room assumes.

Then take the single next move for your stage, and only that. Stage-one firms do not need an operating system yet; they need a commercial plan and shared Projects. Stage-three firms do not need more pilots; they need to extend what is connected.

If you want an outside read on where your firm sits and what the next move is, our AI Readiness Diagnostic scores you fast, and a Discovery Sprint turns the stage into a plan toward the AI Operating System.