How do they compare?
One row per kind of tool. Sources and the day each fact was checked are in its section below.
| Kind of tool | Does well | Stops at | When you change AI | Price | Data kept |
|---|---|---|---|---|---|
| Claude and ChatGPT projects and memory | Files and instructions beside the chat; each person’s memory | One vendor; memory per person; uploads go stale | Projects set up again; memory moved by hand | Claude Team $25 a seat a month; ChatGPT Business per user | Claude in the US; ChatGPT in Europe on Enterprise and Edu only |
| Shared notes for AI | Notes Claude, ChatGPT and other AIs read and write | No tables of figures, no dashboards | Stays: any AI reads them | Hjarni Teams €18 a seat a month | Hjarni in the EU; Basic Memory in the US |
| Dashboard tools | Charts that stay current from connected tools | Figures only; Databox brings its own AI, BitBoard works in SQL | Stays in that tool | Databox from $399 a month; BitBoard free for one, $49 a seat | Databox in the US |
| Workspaces | Docs, databases and dashboard views any AI can reach | People write and update every page | Stays: any AI connects | Notion Business €19.50 a member a month | Notion in the US; the EU on Enterprise |
| Building it yourself | A shared folder your AI reads; screens your AI builds | Someone keeps it in order, by hand | Stays: the files are yours | The tools’ own plans and your time | Where each tool keeps it |
| Omand | One base any AI reads and writes: files, figures, records, a board | Opens to a few companies at a time; runs on the AI plan you already pay for | Stays: the new AI connects to the same base | €149 a month or €1,190 a year per company, without VAT; people without limit | In the EU |
Why not Claude’s or ChatGPT’s own projects and memory?
Because they keep your company’s knowledge inside one AI vendor. A project keeps instructions, files and chats together: Claude shares projects on Team and Enterprise, ChatGPT on every plan. Claude’s Free plan holds up to 5 projects.
Where it stops:
- One vendor. A Claude project does not open in ChatGPT, and memory belongs to each person in Claude and in ChatGPT, not to the company.
- Copies. A project holds uploads, so a new price list goes into every project again.
- Shared pages. In a shared Claude artifact everyone connects their own apps, so two managers can see different numbers. Claude Data shows everyone the same current figures, from a data connector such as a warehouse or a CRM.
- Leaving. Claude’s organization export leaves out the files Claude made.
Microsoft 365 Copilot answers from your Microsoft 365, and Google Workspace keeps data in EU regions from Business Standard; each serves its own suite’s AI.
| Plan | Price | Data kept |
|---|---|---|
| Claude Team | $25 a seat a month, $20 billed yearly, for 2 to 150 people | In the US |
| ChatGPT Business | Per user a month, from 2 users; OpenAI’s page showed no amount | In Europe on Enterprise and Edu only |
With Omand, the knowledge sits in a base your company owns, kept in the EU: any AI connects to it, and a new price list goes in once.
Prices and artifacts checked on 18 November 2026, the rest on 15 November 2026, each on the page linked.
Why not shared notes for AI, such as Hjarni?
They hold text, not figures. Hjarni keeps notes that Claude, ChatGPT, Cursor and any MCP client read and write, with no AI inside: each person brings their own. Teams costs €18 a seat a month (€180 a year), is hosted in the EU at Hetzner and exports everything as Markdown; the free plan holds 25 notes. Hjarni says it serves small technical teams. Basic Memory does the same with servers in the US.
Where it stops: notes hold no tables of figures and draw no dashboards, so the month’s numbers stay in their exports, and someone at ease with AI connects each person.
With Omand, the same base holds the records, the tables every export fills and the board drawn from them.
Hjarni’s prices checked on 18 November 2026, the rest on 15 November 2026.
Why not a dashboard tool such as Databox or BitBoard?
They hold figures, not what people know. They keep charts current from the tools they connect to.
| Tool | How it works | Price | Data kept |
|---|---|---|---|
| Databox | Dashboards from more than 130 tools, built in Databox with its own AI, Genie; its connector sends data in but builds no dashboard | $399 a month billed yearly, $499 monthly, up to 30 data sources and 1,500 AI credits a month; no free plan | In the US |
| BitBoard | Your AI builds dashboards as SQL queries that run again, from a database or apps, or a CSV your AI imports | Free for one person; Teams $49 a seat a month | Not on its pricing page |
Power BI is priced per user; Fathom, for financial reports, by the number of companies.
Where it stops: these tools hold figures, not what people know or the files behind them, and Databox’s AI is a second AI bill beside the one you already pay.
With Omand, each export fills its table, and your own AI describes the page that the Omand app draws, beside the files and decisions behind the numbers, on the AI plan you already pay for.
Prices checked on 18 November 2026, the rest on 15 November 2026.
Why not a workspace such as Notion?
People keep it by hand. Notion holds documents, databases and dashboard views, and any AI connects to it. Business costs €19.50 a member a month and adds Notion’s own agent; data is kept in the US, in the EU on Enterprise. ClickUp, a work suite, prices per user with its AI on credits.
Where it stops: people write and update every page, and figures arrive by hand or through its own connections. A company of 30 pays for 30 members.
With Omand, your AI files what people bring and tell it, every fact with its source and date, at a price per company, not per person.
Notion’s price checked on 18 November 2026, the rest on 15 November 2026.
Why not build it yourself?
You can, with tools you may already pay for, and someone keeps it in order by hand. Keep originals in one Google Drive or SharePoint folder and connect your AI to it: Claude’s directory lists Google Drive and Microsoft 365 among 1,138 connectors. Keep one spreadsheet per kind of export and add each month below the last. For screens, your AI can build them in Airtable, which keeps data in the EU on Enterprise Scale, or in Softr, which keeps it in the EU or the US; Lovable writes each app’s code on its own credits.
Where it stops: someone names the files, adds each export and fixes what breaks, and the AI reads whole files on every question, which uses more of your plan. Nothing checks that a fact has a source.
With Omand, your AI keeps the base in order and every fact keeps its source; nobody tends it by hand.
Claude’s directory checked on 18 November 2026, the rest on 15 November 2026.
What does Omand do that these do not?
Omand holds the parts above in one base: files as they arrived, tables that every new export fills, records of what people know with each fact’s source and date, and a board of pages your AI describes and the Omand app draws. Any AI that takes an MCP connector reads and writes it, each person under their own sign-in; the base is kept in the EU, each company in a database of its own, and exports whole as Markdown, CSV and the original files. When the company changes AI, the new one connects to the same base, and nothing is lost.
It connects once, as a custom connector in the AI you already use. Files go in as they are, and what other tools hold comes in through your AI’s own connectors to them, so there is nothing to install and no integration to build.
Where it stops: it has no AI inside, so every answer runs on your company’s own AI plan; it builds no integrations, so your AI brings what other tools hold and thousands of rows come as an exported file. It costs €149 a month or €1,190 a year per company, not per person, without VAT.
Which AIs connect to Omand lists the plans each AI needs.