What Aqmen is

Aqmen is where analysis for a decision is built — a market sizing, a due diligence, a business case — and where it stays. You work with your own AI agent (Claude, ChatGPT or another), and everything it does lands in a workspace you can open, check and reuse: the documents it read, the data it loaded, the models it built and the deliverable at the end, all connected.

The problem it solves

Analysis done in a chat evaporates. The numbers are right there, but where did each one come from? Which version of the data? What changed since? A week later nobody can say, and the work starts again.

In Aqmen, every number carries its receipt. A figure in a memo knows the chart it comes from; the chart knows its data; the data knows the document and the page it was taken from. When new data arrives, everything that reads it says so.

The pieces: data and artifacts

A project holds two kinds of things, as the app's sidebar shows them.

Data — the evidence, and what is computed from it:

• Sources — the documents behind the work: reports, filings, PDFs, spreadsheets, web pages. Stored as they are, so they can be read again.
• Datasets — clean tables taken from those sources. Every dataset cites where its numbers came from, and every column says what it means: its unit, its scale, its definition.
• Transformations — a new dataset combined from others, the method written down once.
• Charts — a figure from the data, re-run whenever it is opened.

Artifacts — what you build on the data, and what you hand over:

• Spreadsheets for models — Excel-compatible, connected to the data.
• Views for exploring a question, live.
• Docs for memos and findings, decks for presentations, and pages for web reports — downloaded as Word, PowerPoint or a web page, and spreadsheets as Excel.

Every figure in an artifact is read from the data, never typed in. The line from a source to a deliverable is the lineage: open it for anything and see what it reads and what reads it.

Where work lives

• A workspace is usually one client or one engagement. Its data stays in it.
• A project is one piece of work inside a workspace — a market study, a model, a board pack. Its scope — a doc in the project — says the decision and the questions it answers; its notes are where agents keep the job's state, so every session (yours or your agent's) starts where the last one stopped.
• Collections are folders inside a project. Every project has the same skeleton: Scope for the scope doc, one collection per workstream, and Deliverables.

When you sign up you get a space of your own. Start with your own work straight away, or learn first with a guided example in a Playground — made when you choose it, and yours to delete when you're done.

Working with your agent

You can do everything in the app. Most people let their agent do the heavy lifting: you say what you need, it researches, loads the data with its sources, builds the model, writes the memo — and you read, check and steer. The app is where you see what it did, open any figure to its source, and change what you disagree with.

Your agent doesn't start from zero each time: it reads the project's notes, the data and what was concluded so far before it begins.

What it is not

Aqmen is not a place to clean messy data. Data arrives in a workspace already clean — your agent cleans it before loading it — so that every table can be trusted and combined. Transformations derive new tables from clean ones; they don't repair broken ones.

Next

Connect your agent — a minute, once — then, if you'd like a guided start, follow Your first project: an example project with its data already in, a figure and a memo in five minutes.