Evo Researcher is an AI-powered research and collaboration tool built into the Evo platform. Its a project-based AI chat with file uploads, connectors, templates and team collaboration, all aware of the wider Access Evo ecosystem.

It's built for work that involves reading, analysing or producing documents, where the context matters as much as the question. You give it your files and instructions upfront, and it works within that context across every conversation in the project.

This guide focuses on self-service use of Evo Researcher for finance teams, including what it is, the key components you'll come across, and how to get started.

Researcher lives inside the Access Evo mobile app, as well as in your desktop web browser. It is free to download and works on both iOS and Android devices.

When you first open the app, you will be asked whether to allow push notifications. You can change this preference at any time in your device Settings.

Everything in Researcher lives inside a project. Once you're in a project, the workspace breaks down into three areas:

(New Researcher template)

A role-based onboarding template for HR Directors, People Managers, School Leaders, HR, and Payroll Administrators in education. It introduces Researcher through practical HR, workforce, absence and pay tasks, guiding you step by step. "Show me" examples use live data from your Education People HR system.

(New Researcher template)

A ready-to-use template for querying Education People HR data. Surfaces workforce reports, gender pay gap analysis, absence analytics, payslips, leave, and personal details, all from live data via the Education People connector. Copy and customise it for your own projects, or use it as-is to explore what's possible.

There are two ways to access the templates:

Before the AI can give you HR-specific answers, it needs your context. There are three ways to give it that context:

(Project setup)

Upload documents to your project and the AI can reference them across every chat, without you needing to re-explain the context each time. Files are indexed so even long documents can be searched without filling up the conversation.
​

(Single message)

For a one-off file relevant to a single question, attach it directly in the chat using the paperclip icon. It's used for that message only and isn't added to the project's permanent uploads.

(Live data)

Instead of uploading a static file, connect to a live data source from Evo products or other tools, so the AI works from current information rather than a snapshot that can go out of date.

On top of the general Uploads and Connectors described above, the following HR-specific connectors and templates are planned for Researcher.

A direct connector letting Researcher query Access Education People data, including data from the People Insights Hub.

A direct connector letting Researcher query Access SCR Manager

Chat consent is a pop-up that asks for your permission before Researcher changes something through a connector, such as creating, updating or deleting a record. Researcher only asks before a change, not when it reads information.

When the pop-up appears, you can choose:

Permissions are set per action, not per connector. Approving one action, such as updating a record, does not cover a different action on the same connector, so each new action asks once. Workflows that use many different actions will prompt a few times at first, then settle as the common ones are approved. Permissions also apply per project and per person, so each teammate approves separately.

To review or remove anything you have allowed, go to User settings, then Chat consent.

Everything in Researcher lives inside a project. A project keeps a set of instructions, uploads, connectors and chats together, so the AI has consistent context for a particular piece of work.

A conversational interface where you ask questions and get answers, automatically informed by everything already in the project; no need to remind the AI what it's working on each time.

(Consistency)

A standing brief you set once per project to control how the AI responds; for example, always opening with a status summary, or always using the same section order. Without instructions, the shape of the answer can vary each time you ask a similar question.

(Reusable playbooks)

Reusable instruction sets that an admin publishes once, which any user can attach to a project in one click; useful when your team follows the same playbook across many pieces of work.

(Consistency)

When the AI produces something, you want to send, share or keep, a report, an email, a summary, it can generate it as an Artifact. Artifacts are kept separate from your uploaded documents, and can be previewed, downloaded, or copied as rich text so formatting doesn't break when pasted elsewhere.

(Scaling up)

A pre-configured project setup, instructions, uploads and connectors already in place that lets you or your team spin up a new project for a repeatable piece of work in one click, rather than setting it up from scratch each time. The HR and safeguarding team templates in Section 5 are examples of this.

(Working together)

Add colleagues to a project so they see the same chats, uploads, instructions and artifacts; useful when a piece of HR or safeguarding work is owned jointly.

Evo administrators will need to enable individual users access to Researcher, they can do this by:

No, use the same credentials you already use for your HR system.

Yes. Everything in this guide is available on the Evo mobile app as well as in your browser.

Notifications appear in your Evo Feed; ones that need you to act sit in your Action Required list too. If you have the mobile app, you'll also get push notifications, check your mobile push settings if you're missing them.

For data questions (your projects, uploads and artifacts); your HR team.
For technical issues with the navigator screen itself; [email protected].

Uploads are for documents you want the AI to reference across the whole project. Attachments (the paperclip icon in chat) are for a single message only. Connectors give the AI live data from a system instead of a file that can go out of date.

No. Without instructions, you can still chat with the AI in a project. Instructions simply make repeated tasks come back in the same shape every time, useful once more than one person is using the same project.

Most often because Instructions weren't set, or because the question was asked in a new chat that hadn't picked up the saved Instructions yet. Save your Instructions in the project and start a new chat to get a consistent shape going forward.

Yes, update them in the Instructions tab and save. Start a new chat afterwards so you get a clean response built on the updated brief; chats already open may not pick up the change straight away.

Artifacts are treated as project outputs, they sit in the Artifacts tab, and can be previewed, downloaded or copied as rich text. A normal chat response is conversational only, unless you ask the AI to save it as an Artifact.

Only members with access to Researcher and specifically collaborators you have added into the project can see the content.

No, while the AI interacts with and retrieves answers from your internal system data or documents when prompted (such as searching contracts or checking records), it acts dynamically on that data without absorbing it into a permanent public or shared training model

No. You do not need technical skills or coding knowledge to use Access Evo Researcher or its connected systems