As an Early Access participant, you are testing the Rally Agent before it goes to the world, and the things you run into, the feedback you share, and the questions you ask are what make it better. So: thank you. We mean it.

This article tells you exactly what needs to be in place in your workspace for it to work well. 

Interact with the Agent by clicking on your profile picture in the bottom left corner of your Rally workspace, then entering Agent mode! 

A note on approvals with the Rally Agent in early access

In Early Access, every action the agent takes comes to you for approval before it executes, recruitment proposals, outreach, incentive sends, study closes, all of it. This is by design.

The model is approve-now, automate-later: you see what the agent proposes, build intuition for how it works, and flag anything that doesn't look right. Your feedback on individual decisions is some of the most valuable signal we get as we build toward fuller automation.

Steps to set your workspace up for success with the Rally Agent 

The Agent works with what is already in your Rally workspace. The more of this you have in place before you start, the more useful it is from session one.

Create study templates

The Agent searches your templates and proposes the best match when you ask it to start a study. 

Create at least one interview study template before your first Agent session. Each template should include a descriptive name that signals what the study is ("30-min onboarding exit interview," not "Template A"), the session duration, a default incentive amount, any custom fields your team captures, and screener questions if you use them. The Agent matches on the template name and structure, so the clearer and more descriptive each one is, the better the match.

Populate participant properties

The Agent sources participants based on properties stored in your participant CRM. Unstructured people search is not available in EA and all audience sourcing works off defined property criteria. If those properties are empty or inconsistently filled in, the Agent has nothing to filter on and cannot surface a qualified audience. Add property descriptions to help the agent understand the purpose of each property. 

Go into your participant CRM and fill in the properties your team actually recruits on. You do not need every field filled for every participant. You need the criteria you filter by most often to be consistently populated across your panel. Common examples: role, seniority, company size, industry, product tier, and region.

If a property you recruit on does not exist in your workspace yet, create it before you build your first Agent-assisted study.

Build saved audience views

The Agent can find participants by property criteria on its own, but a saved audience view speeds things up. When you ask the Agent to source an audience, a named and filtered view gives it a clear, ready-to-go starting point instead of building from scratch each time.

Save the audience segments you come back to repeatedly. Examples: "Enterprise customers," "Recent onboarding completions," "Active B2B users on the free tier." Each saved view should have a name that tells anyone (including the Agent) exactly who is in it without having to open it, and it should be filtered on the participant properties you filled in above.

Configure your governance settings

Every frequency cap, consent requirement, and permission rule you have set up in Rally is what the Agent enforces automatically. It will not contact a participant more often than your rules allow. It will not skip consent steps you have required. The guardrails you set are the only guardrails it operates within.

Before your first Agent-driven recruitment run, confirm these things:

• Engagement rules are set. These control how often any single participant can be contacted and within what time window. 

• User permissions are current. During EA, access to the Agent is controlled. Make sure the right people on your team are enabled.

Configure your incentives

The Agent triggers incentive payouts 24 hours after a completed session and routes them to you for approval. Without incentives configured in Rally, this capability simply will not work.

Before your first Agent-assisted study, connect and fund your incentives in Rally's incentive settings and confirm the denominations your team uses. When the Agent surfaces an incentive for approval, you review it, approve it, and it sends.

Connect your calendar and set availability

The Agent can schedule, reschedule, and cancel interviews, but only if your calendar is connected. Without it, the Agent can propose times but cannot confirm bookings.

Connect Google Calendar or Outlook in Rally's scheduling settings, and set the availability windows you want to offer for sessions. Keep your availability current. Participants book directly against it.

Set a participant target on every study

The Agent monitors session completions against your quota and surfaces the study close when the target is hit. Without a quota set, it has no signal for when a study is done.

Every study your team builds – whether you built it or the Agent did – should have a participant target set before it goes live. This keeps recruitment from running indefinitely and gives the Agent a clear finish line to work toward.