Skip to main content

Frequently asked questions

B
Written by Bianca Manase

Find answers to common questions about getting access, chats and projects, connectors, artifacts, skills, templates and troubleshooting in Access Evo Researcher.

This article answers the questions we hear most often about Access Evo Researcher. Use the side navigation to jump to the section you need.


Getting started

How do I get access to Access Evo Researcher?

An organisation administrator turns on Access Evo Researcher in the Evo Platform Administrator tool. After that, each user is assigned either the User or Admin role. Access Evo Researcher then appears in the Evo navigation menu automatically.

Does Access Evo Researcher work on mobile?

Access Evo Researcher is optimised for desktop. A mobile-friendly experience is not available in the current release.

Which AI model does Access Evo Researcher use?

Access Evo Researcher has two modes: Standard and Advanced. Standard mode uses Haiku 5.5, and Advanced mode uses Sonnet 5.5.

You're in Standard mode by default. To use Advanced mode, select Advanced in the chatbox.

Where do I build an AI agent or assistant?

Use the Assistant Builder app to build AI agents. Don't try to build agents within Access Evo Researcher.

What does Access Evo Researcher do if I don't upload anything or connect a data source?

Without uploads or connected data sources, Access Evo Researcher works like a general AI chat. It doesn't know about your customers, projects or organisation until you upload a document or connect a data source. It can still search the web for general information.

Can I continue a Quick Asks conversation in Access Evo Researcher?

Yes. From Quick Asks in Navigator, you can open the conversation in Access Evo Researcher. For details on how projects and templates appear in Quick Asks, and what base and advanced users can see, read the Researcher in Quick Asks article.


Chats and projects

Does Access Evo Researcher remember anything from a previous chat?

Not across chats. A new chat only sees the project's instructions, uploaded documents and saved artifacts. It doesn't see the message history of another chat, and chats in different projects can't reference each other.

If you want something from one chat to be available to every chat in the project, ask Access Evo Researcher to save it as an artifact.

Why should I start a new chat instead of continuing my current one?

Every message resends the full conversation history to the AI. A long chat costs more tokens for each question and can mix unrelated topics into the same context. Start a new chat when you move to a different task, and stay in the same chat for follow-ups on the same topic.

If you need something from an old chat, ask Access Evo Researcher to summarise it and paste the summary into the new chat, or save the key information as an artifact.

Can I attach a file to a chat?

Yes. Select the attachment icon in the message box. The AI uses the file to answer that message. If you want the AI to refer to a file throughout a longer piece of work, add it to a project as an upload instead.


Uploads and files

What file types can I upload to a project?

You can upload .txt, .json, .csv, .md, .pdf, .docx, .xlsx, .xls, .pptx, .html, .xml, .rtf, .yaml, .yml and .log files, up to 30 MB each. Excel and Word files work directly, so you don't need to convert them first.

Video files aren't supported. Stick to documents, spreadsheets and images.


Connectors

Can I use connectors in a single chat?

Yes. Default connectors are added to every new chat and project, so you don't need to add them each time. You can manage your default connectors in your user settings, under Connectors.

You can also add connectors directly to a single chat in chat mode, without creating a template. Connectors can also be included in templates.

What types of connector are available?

Access Evo Researcher supports four types of connector:

  • Evo product connectors give the AI live data from other Access products.

  • Third-party connectors connect to tools such as OneDrive using OAuth.

  • Agent Builder agents can be added as connectors.

  • Custom MCP servers can be connected to your project.

The Evo product connectors you see depend on your region and on which products your organisation has turned on. To see which connectors are available to you, check the Connectors tab in your project.

How do I add a custom MCP server?

Open your project and select the Connectors tab. Select + Add connector, select Custom, then paste the MCP server URL and add any authentication headers the server requires.

How does the AI use a connector?

Adding a connector gives the AI a tool it can call when a message needs that data. The AI decides when to use it based on what you ask. Each tool call is visible in the chat response, so you can see what was queried.

🤓Tip: Only add the connectors your project needs. Each connector adds to the AI's decisions and can reduce the accuracy of responses.

Are connectors shared between projects?

No. Connectors are set up per project, so every new project needs its own connector setup. Projects created from a template get their connectors automatically. If you repeat the same connector setup across projects, convert the project into a template.

Can a collaborator use my connected account?

No. Third-party connectors use per-user OAuth. Each collaborator authorises their own account the first time they use the connector. Templates set which connector is attached, but each user supplies their own credentials.

How does Access Evo Researcher send an email?

There are two ways to send email:

  • Evo Comms connector: Sends the email straight away when you tell it to send. Confirm the recipient and final wording before you say send.

  • Evo Emailer action: Opens a pre-filled email panel with From, To, Subject and Body fields. Check and edit the email, then select Send yourself.


Artifacts

What types of artifact can Access Evo Researcher generate?

Access Evo Researcher can generate HTML pages, Markdown documents, PDFs, Excel spreadsheets, PowerPoint presentations and charts. Word documents (.docx) can't be generated as artifacts.

Can I change an artifact's format after it's created?

No. An artifact's type is fixed once it's created. To get the same content in a different format, ask the AI to produce it again as the new type.

How do I get the AI to create an artifact?

If the AI replies in the chat instead of creating an artifact, ask it explicitly, for example: "Save that as an artifact I can preview and download." The artifact then appears in the Preview and Artifacts tabs.

Can I paste an artifact into Outlook or Word?

Yes. Use the Copy action rather than selecting the text in the preview. Copy keeps the formatting. Selecting text from the preview can bring in raw Markdown syntax, which shows up as clutter in Outlook or Word.

How do I share an artifact?

Artifacts require sign-in, so a direct link won't work for someone who isn't logged in. You can either:

  • Download the artifact and share the file.

  • Add the person to the project as a read-only collaborator, so they can view its chats and artifacts in Access Evo Researcher.

Does Access Evo Researcher keep artifact versions?

No. Editing an artifact replaces it in place. If you want to keep the original, say so in your request, for example: "Don't overwrite this. Save the edit as a new artifact named version 2."


Skills and templates

What is the difference between instructions, a skill and a template?

  • Instructions are guidance you write for one project in the project's Instructions tab.

  • A skill is a reusable block of instructions. Once a skill is shared with the organisation, any user can attach it to their projects.

  • A template is a full project blueprint. It can include instructions, uploads, connectors and actions.

A typical approach is to write instructions for one project, publish them as a skill once you use them repeatedly, and fold them into a template if the workflow also needs the same uploads and connectors every time.

Who can create skills and templates?

Any user can create personal skills for their own use. Organisation administrators can create and publish skills and templates for the whole organisation.

Access and roles

What can an Admin do that a User can't?

Admins can create and manage organisation-level templates and skills, manage users and configure organisation-wide settings. Users can create and manage their own projects, upload files, chat with the AI and collaborate on shared projects.

Can I hide a system template my organisation doesn't need?

Yes. Organisation admins can hide system templates, which are cross-organisation templates set by the product team, without deleting them.

Can a collaborator see data from my connectors?

Yes. A collaborator can see everything in the project, including chats, artifacts, uploads and any data already pulled in through your connectors, even if they don't have their own access to those connectors. Access is checked when data is pulled into the project, not when a collaborator views it. Think about what you bring into a project before you add collaborators.

To give someone view-only access, add them as a read-only collaborator. They can see the project's chats and artifacts but can't edit or add to the project.


Tokens

How can I use fewer tokens?

Prompts

  • Be specific and direct. Vague or wordy prompts need extra rounds and use more tokens.

  • Bundle your requests into one message where you can.

  • Don't repeat context the AI already has. If you uploaded a file earlier, don't paste its content again.

Instructions

  • Keep project instructions short and focused on how the AI should behave. The instructions are resent with every request in the project.

  • Link to a file instead of including examples or data in the instructions.

Connectors

  • Only add the connectors your project needs. The details of every connected tool are sent with every request, whether it's used or not.

Uploads

  • Upload each file once per chat. Don't re-upload the same file.

  • Use smaller, focused files where you can.

  • Put documents needed across many chats in project instructions or a data space instead of re-uploading them.

Artifacts

  • Ask for Markdown before HTML, because HTML uses more tokens. Only request HTML where you need it.

  • Ask for updates to an existing artifact rather than a new one. For long refinements, continue in a new chat.

Did this answer your question?