Connectors give the AI access to live data and systems beyond what you've uploaded to a project. Where an uploaded file is a snapshot, a connector is a live link — so the AI can fetch current information, and in many cases take action, rather than working from something that may already be out of date.
Why use a connector?
Uploads work well when content is stable — a contract, a process guide, a set of notes. They stop working well when the content changes regularly.
Connectors solve this. Ask "how many open support tickets does this customer have?" against an uploaded spreadsheet and you get last month's answer. Ask the same question through a connected helpdesk and the AI fetches the live count.
Connectors can also write back to their source, not just read from it — so the AI can update a record, create a task, or log a note, not just retrieve information.
The four types of connectors
All connectors are added from the same place — the Connectors tab inside a project — and the setup wizard guides you through whichever type you're adding.
Type | What it is | Typical example |
Evo product connector | A route into another Evo product, exposed as a tool the AI can call | The CRM, the HRIS, Data Spaces |
Third-party connector | A pre-built integration with an external SaaS, authorised per user via OAuth | OneDrive, Asana, ADO |
Assistant Builder assistants | An AI assistants built in Access's Assistant Builder, surfaced inside Researcher as a tool | A bespoke triage agent your team has built |
Custom (MCP server) | A user-defined connection to any MCP-compliant server | An internal service you've stood up that speaks MCP |
Add them all from the same + Add connector wizard. Pick the type on the first step and the wizard branches accordingly.
What's available
Evo product connectors include Data Spaces and a range of internal Evo products such as CRM, HR, and recruitment tools. Which ones are available depends on your organisation's setup and region.
Third-party connectors include OneDrive, with more on the way. Each person who uses a third-party connector in a project needs to sign in with their own account the first time — their access is kept separate from everyone else's.
Assistant Builder assistants are fetched automatically from your organisation's Agent Builder. Any agent your team has published there will appear as an option.
Custom connectors let your technical team connect Researcher to any internal or external service that supports the MCP standard. Your IT or development team would set these up.
Setting one up
1. Open a project and click the Connectors tab (the plug icon on the right-hand side).
2. Click + Add connector.
3. Choose the type — Evo product, third-party, Assistant Builder assistant, or custom.
4. Follow the setup steps, which vary by type:
- Evo product — choose which capabilities the AI is allowed to use.
- Third-party — sign in with your account.
- Assistant Builder assistant — pick from the assistants your organisation has published.
- Custom — your IT team provides the connection details.
5. Save. The connector is now available to the AI in this project.
Each connector shows a status badge so you can see at a glance whether it's connected and working.
How the AI uses a connector
The AI doesn't monitor connected systems in the background. Instead, each connector gives the AI a set of actions it can choose to take when your message calls for it.
For example:
- You ask: "What's open for this customer in the helpdesk?"
- The AI recognises it needs to check the helpdesk, calls the right action, gets the result, and summarises it for you.
- The chat shows you what was retrieved, so you always know where the answer came from.
A note on keeping things focused: every connector you add is another option the AI has to consider on every message. A project with many connectors gives the AI more to weigh up, which can lead to more mistakes. Add the connectors the work actually needs — leave others out.
Privacy and access
Different connector types handle credentials differently, which matters when a project is shared with collaborators.
- Evo product connectors — handled automatically through the platform. Each person's access reflects their own permissions in the Evo platform.
- Third-party connectors — each collaborator signs in with their own account. The AI acts on behalf of whoever is currently using the project, not the person who set the connector up.
- Assistant Builder assistants — authentication is handled through the Agent Builder integration.
- Custom connectors — depending on how the connector was configured, it may use a single shared credential or require each person to authenticate individually. If you're unsure which applies, check with whoever set it up.