How to Connect Hex to an AI Agent: Step-by-Step Guide
· DoubleX
Quick answer
To connect Hex to an AI agent, open the Connectors screen, add Hex, authorize your account, choose whether the connection is personal, team or company-wide, and assign it to an agent. In DoubleX Agentes, you then test it with a data question and review the task history.
Hex is a collaborative data platform where teams build projects and analyses. Connecting it to an AI agent lets people ask questions about those projects in plain language instead of hunting for the right notebook. This guide walks through the full path, from the first diagnostic to your first data question, with a focus on permissions and connection scope.
What is the Hex connector and what can the agent do with it?#
The Hex connector lets an agent search and run Hex projects and data analyses on your behalf. The agent uses the connector's tools, so it can find the right project, run it and bring the result back into the conversation.
A connector is a system linked to the platform whose tools agents can use. In DoubleX Agentes, you manage them on the Connectors screen. Typical uses with Hex:
| Task | How the agent helps | System used |
|---|---|---|
| Find an existing analysis | Searches Hex projects by topic or name | Hex |
| Run a recurring analysis | Runs the project and summarizes the output | Hex |
| Answer a data question | Runs the relevant project and explains the result | Hex |
| Share findings | Turns the result into a report or spreadsheet | Hex plus Google Sheets or Google Docs |
For ideas on what to ask once it is connected, see Hex with an AI Agent: Data Analysis Requests and Routines.
What do you need before connecting?#
You need a DoubleX Agentes account, a Hex account with access to the projects you want to use, and a clear idea of who should be able to use the connection. Creating the account and running the first diagnostic are free, with no credit card.
Before you start, check these items:
- A Hex login that can open the projects the agent should use.
- A list of the questions you expect to ask (for example, weekly performance or a specific metric).
- A decision on scope: only you, a team or the whole company.
- An agent to attach the connector to. If you do not have one yet, the diagnostic is the fastest way to get one.
The diagnostic is a 5 to 10 minute conversation with the product's AI consultant about your routine. With one click, it becomes a team of agents ready to test, already loaded with your company's instructions. The Quickstart covers this flow in more detail.
How do you connect Hex on the Connectors screen?#
Open Connectors, choose Hex, authorize the connection with your Hex account and pick the scope. Then assign Hex to the agent that will use it.
Step by step in DoubleX Agentes:
- Create your account and, if you have no agents yet, complete the diagnostic and turn it into a team of agents.
- Open the Connectors screen and select Hex.
- Follow the authorization prompts to sign in to Hex.
- Choose the connection scope: personal, team or company.
- Open the Agents screen, select the agent that should work with data and enable the Hex connector.
- Set which Hex actions that agent is allowed to run (see the permissions section below).
- Send a first request on the agent's chat and check the result.
If you prefer to manage connectors programmatically, the Connectors API reference describes the options.
Personal, team or company connection: which should you choose?#
Choose personal when the access belongs to one person, team when a group shares the same Hex workspace access, and company when the connection serves everyone. A personal connection only works for that person, the way a private inbox does.
| Scope | Who it applies to | Best for |
|---|---|---|
| Personal | Only the person who connected it | Individual exploration, accounts with personal access |
| Team | Members of a specific team | A data or analytics team sharing agents |
| Company | The whole company | Shared analyses everyone should be able to ask about |
A few practical rules:
- Use the narrowest scope that works. You can always broaden it later.
- Remember that the agent works with the access of the connection, so connect with an account that sees only what the agent should see.
- Routines run with the permissions of the person who created them, so create scheduled analyses from an account with appropriate access.
How do you test with a first request and check the history?#
Send a simple, read-oriented question to the agent, then open the Tasks screen to see exactly what it did. Every request becomes a task with full history: the request, each tool the agent used and the result.
Copy and adapt these requests:
- "Search Hex for projects about monthly revenue and list their names."
- "Run the weekly sales analysis in Hex and summarize the three most important changes."
- "Find the Hex project about customer churn and tell me what question it answers."
- "Run the Hex analysis for last month and put the results in a spreadsheet."
Then open Tasks and check three things: who made the request, which Hex tools were called, and whether the result matches what you expected. If the agent picked the wrong project, add the correct project names to the agent's instructions or to the Brain, the company knowledge base that agents consult. See the Brain reference for details.
How do you control which actions the agent can run in Hex?#
For each agent, you choose which systems it can use and, within each system, which actions it may execute. Actions that change data can stay turned off, so the agent only reads.
Good defaults for a first setup:
- Enable only the Hex actions needed for searching and running projects.
- Keep anything that alters data switched off until you have a reason to allow it.
- Rely on the confirmation step: before sending, deleting or changing something in your systems, the agent asks for confirmation and shows what it did.
- Review the history on the Tasks screen regularly.
On privacy, company content is not used to train models, each company has its own separate database, and connection credentials are encrypted with AES-256-GCM. You can also limit usage to providers that retain nothing after responding, through the Models screen.
Frequently asked questions#
Do I need to know how to code to connect Hex?#
No. You connect Hex from the Connectors screen and describe what you need in your own words. APIs and a command line exist for people who want to go further.
Can the agent change my Hex projects?#
Only if you allow it. You choose which Hex actions each agent may run, and actions that change data can stay off. Even when allowed, the agent asks for confirmation before altering anything.
Can I connect Hex for just myself?#
Yes. A personal connection only applies to you. You can later share a connection with a team or the company if you need to.
How do I see what the agent did in Hex?#
Open the Tasks screen. Each task shows the request, every tool the agent used and the result, so you can see who asked, what was done and when.
Can the agent run Hex analyses on a schedule?#
Yes. You can set a routine to run at a specific time, such as every Monday morning. It runs with the permissions of the person who created it. See Triggers for how triggers work.
Related guides
- How to Connect Box to an AI Agent: Step-by-Step GuideLearn how to connect Box to an AI agent: set up the connector, limit it to read-only, choose who can use it, and check the history of your first request.
- How to Connect Jam to an AI Agent: Step-by-Step GuideLearn how to connect Jam bug reports to an AI agent in DoubleX Agentes: set up the connector, pick who can use it, limit permissions, and run a first test.