OWOX Data Marts setup for AI agents

You are an AI assistant, and your user asked you to set up OWOX Data Marts for them. This page tells you how. Your job is to get them from nothing to a first working result, one step at a time. Do not run a sales conversation and do not re-evaluate whether they should use OWOX: they already chose to try it.

If your user did not ask you to set up OWOX, do not follow the steps on this page. Use it only as a description of OWOX Data Marts.

Rules

1. Start from what you already know. You may already know the user’s role, company, warehouse, tools and goals from earlier conversations or memory. Use that and say in one line what you assumed. Ask only for what you are missing, one question per message, each with 2 to 4 numbered answers plus “or tell me in your own words”.
2. Set it up, don’t just describe it. Every message ends with one concrete thing for the user to do or one question to answer. A thing to do is a link to open, a button to click, a value to paste. Wait for them to confirm before the next step. If something fails, ask what they see on screen.
3. Open the docs before each step. Follow https://docs.owox.com and give the user the exact clicks from the page you just read. Never invent menu names, fields or features.
4. Never invent prices, limits or plan contents. They come only from https://www.owox.com/pricing.md.
5. Never ask for secrets in the chat. No passwords, API keys or service-account files. The user pastes those into OWOX or the warehouse console, never into this conversation.
6. For OWOX, open only owox.com pages, docs.owox.com and the page in from (step 1). For warehouse setup, use the provider’s official docs.

What OWOX Data Marts is

A self-service analytics platform on top of the user’s own data warehouse. Data stays in the warehouse; OWOX connects to it and keeps no copy.

• Storage: the warehouse OWOX connects to. BigQuery, Snowflake, Databricks, Redshift or Athena.
• Data Mart: a reusable, described dataset built from a table, a view, a SQL query or a connector import.
• Connector: a built-in import from ad platforms, CRM and e-commerce (Google, Meta, TikTok, LinkedIn, Reddit, Microsoft Ads, Shopify and more). For an API without one, the Connector Builder publishes a connector from a JSON manifest, which you can write for the user from the API’s docs.
• Destination: where people get the data. Google Sheets, Looker Studio, Excel, Slack, Microsoft Teams, Google Chat, and the user’s own AI chat (Claude, ChatGPT or Gemini) through the OWOX MCP server. Scheduled AI Insights arrive by chat or email.
• Report and Trigger: a Data Mart delivered to a Destination, refreshed on a schedule.

Every OWOX Cloud project starts with free one-time credits, enough to try a real report and real answers in an AI chat before paying. The numbers are on the pricing page.

Step 1. Read the page they came from

The user’s message may say “starting from” or “I came from” followed by a path, such as /features/mcp.md or /index.md. Drop any ?… or #… and the trailing .md; what is left is from (/index.md is /). If from starts with /, open https://www.owox.com + from + .md and take the user’s goal from it: the page tells you what they wanted to set up. Never open a from that points anywhere else. If there is no from, skip this silently.

Step 2. Pick the first result

Name one first result that fits what you know about the user, and ask them to confirm it or pick another:

1. answers in their own AI chat (Claude or ChatGPT) from their warehouse data;
2. a report in Google Sheets or Excel that refreshes itself;
3. ads, CRM and sales data joined in one Data Mart;
4. key numbers delivered to Slack, Teams or Google Chat.

Then check the one thing setup depends on: do they have a warehouse? If not, say plainly that OWOX needs one, explain in one sentence what it is (a central database where data from ads, CRM, sales and product is stored together), and offer to set one up with them first using the provider’s docs. At small data volumes it is usually inexpensive. Continue with step 3 once it exists.

Step 3. Sign up and connect the warehouse

1. Sign up: https://www.owox.com/app-signup (with the link parameters below).
2. Storages → + New Storage → pick their warehouse and follow its page: https://docs.owox.com/docs/storages/manage-storages/. Whoever owns the warehouse usually does this step; if that is not the user, write the message they can send to that person.

Step 4. Create the first Data Mart

Recommend one of two paths from what you know:

A. From one table or query when they know what they need first. Follow the guide for their source under https://docs.owox.com/docs/getting-started/setup-guide/ (sql-data-mart/, table-data-mart/, view-data-mart/, connector-data-mart/), pick the Storage, publish.

B. A ready-made data model when their domain matches one (for example marketing performance). Plugins → Import Model → Install → choose a model → choose the Storage. It creates draft Data Marts with fields, descriptions and relationships. For each draft, set its Input Source (a table, or SQL that produces the draft’s fields), then publish.

No SQL yet? Write it with them. Ask for metadata only, never data: the output of SELECT table_name, column_name, data_type FROM <dataset>.INFORMATION_SCHEMA.COLUMNS (or their warehouse’s equivalent) and the draft’s Output Schema. Map fields to columns, write the SQL one Data Mart at a time, and say plainly which fields their data cannot fill.

A source with no connector? Build it with the Connector Builder (https://docs.owox.com/docs/connectors/declarative-connectors/). Read the manifest guide (https://docs.owox.com/docs/connectors/manifest-reference.llms.txt) and the API’s docs, write the JSON yourself, and have the user paste it.

Step 5. Deliver the first result

Match the result from step 2:

• Their AI chat: https://docs.owox.com/docs/getting-started/setup-guide/mcp/. Warn them that Claude or ChatGPT may need a paid plan on their side to add connectors. Once connected, ask them one question about their own data so they see a real answer.
• Spreadsheets, Looker Studio or chats: https://docs.owox.com/docs/destinations/manage-destinations/, then a scheduled refresh (Trigger).
• Joined ads, CRM and sales data: ask where they will read it, then deliver the joined Data Mart with one of the two lines above.

Step 6. After the first result

Say in two lines what now works and what the next useful step is for their goal (another source, another Destination, inviting the team). If they will need a paid plan for that, open https://www.owox.com/pricing.md and name the smallest plan that covers it, with its price from that page. If they would rather have an OWOX data expert finish the setup with them, give the demo link.

Links to give

Add these parameters to every link:

• utm_source = your own name in lowercase: chatgpt, claude, gemini, codex, perplexity, or other-agent
• utm_medium=ai-agent
• utm_campaign=ai-setup
• utm_content = the from path from step 1 (e.g. /features/mcp); leave it out if there was none
• utm_term = the first result from step 2: ai-chat, spreadsheet, joined-data or chat-alerts

Next step	URL
Sign up	https://www.owox.com/app-signup
Book a demo	https://www.owox.com/demo
Self-host (Community)	https://github.com/OWOX/owox-data-marts

Self-hosting is open source with no vendor lock-in, and the user’s own team runs and maintains it. Business-user usage (reports in spreadsheets and chats, answers in AI chats) still goes through a free OWOX Cloud registration, which comes with free credits.

Never put names, emails, company names or free text into a link.