FAQs

How Do I Add Members to My Organization?

Go to Organization Members
Select Add Member (or Invite Member) to open the invite form
Enter their email address and choose their role
Click Invite to send the invitation

By default, invites are limited to your organization’s email domain. If you need to invite someone with a different email domain, reach out to the JustAI team.

Templates and Variants

What are Attributes and why do they matter?

Attributes define the segments of your audience that will receive different variants. They must be added during test creation and cannot be edited afterward. Attributes are tied to your data pipeline—contact your CSM if you need to add a new one to ensure it’s properly mapped.

How do I create and manage variants in Studio?

From the Preview mode:

Click the + icon to generate new variants
Provide additional context if desired
This will open the chat and AI will generate suggestions
Click Create to move preferred variants to Draft
Review and edit draft variants as needed

From the Compact mode:

Open the chat to ask for new variants
You can click on “Create Variants” or write in the request with additional context
Click Create to move preferred variants to Draft
Review and edit draft variants as needed

What actions can I take on a variant?

From the Preview mode:

Approve: Click the green button
In the kabab menu:
- Reject
- Delete
- Duplicate
Edit: Click the variant, make changes in the variable directly, then click Save
Remix with AI: Use the record icon to prompt a new version
More actions: Send Proof, Duplicate, Revert to Draft, Delete, Change Attribute Plan

From the Compact mode:

Approve: Click the green button
Dismiss: x button, this moves this variant into Draft where you can edit, approve, or reject
Edit: Click the pencil to make changes in the variable directly or remix

IMPORTANT NOTE: Don’t forget to hit SAVE.

Can we edit the variant?

Yes. You can edit all parts of the variants directly in the JustAI platform. Be sure to save your changes before navigating away.

From the Preview mode: Click the variant, make changes in the variable directly, then click Save
From the Compact mode: Edit: Click the pencil to make changes in the variable directly or remix

Can we review the full email message for each variation, or only individual elements?

Each variant includes all the variables you’re testing, so this can be subject line, preheader, ct, body/HTML, title, message, etc.

Are the variations tested one at a time, or all at once? Can we control that?

Variants are tested in parallel (not sequentially). Our AI manages how many variants are active and when new ones are introduced. If you prefer to change this amount contact us to make those updates.

If we set up a test with approved variants for only a few segments (called “attributes” in JustAI), what happens to users in segments without assigned variants?

Any segment without an assigned and approved variant will automatically receive the control version of the message.

How do I change the status of a template in JustAI?

To update a template’s status:

Open the template.
Click Configure in the top menu.
On the General Settings page, find the Status section.
Choose the new status from the dropdown menu.
Click Save Changes to apply the update.

Available statuses:

Draft – Default for new templates; used while you’re still editing and reviewing variants.
Active – Variants are approved and ready to go in JustAI (does not mean live in your ESP yet).
Shipped – The test has ended and been finalized.
Archived – The test has ended and is stored for reference.
Deleted – The test will be permanently removed.

Testing and Optimization

How long should we run a personalization test to get reliable results?

We recommend aiming for 10,000 total sends and at least 1,000 sends per variant to gather meaningful results. But this depends can vary based on the frequency of sends and your campaign.

For example, if you’re testing 15 combinations, you’ll want to plan for at least 15,000 sends.

What is Autotuning?

Autotuning is our system’s way of continuously optimizing your campaign. It monitors performance, recommends archiving underperformers, and suggests new draft variants based on what’s working—all while keeping you in control.

To learn more: Autotuning: Smarter, Ongoing Optimization for Your Campaigns

When does Autotuning take action on a variant?

Autotuning begins evaluating a variant once it reaches 1,000 sends. At that point, if the variant is statistically underperforming compared to others in the test, the system may recommend archiving it or automatically archive it depending on your settings.

In some cases, our system may take action before the 1,000-send threshold, especially if early performance signals are strong and the send volume is lower. This helps maintain test momentum when it’s clear that a variant isn’t performing well.

To keep your tests running smoothly, we recommend:

Approving multiple variants in advance so autotuning has options
Monitoring performance thresholds and adjusting them if needed (reach out to the JustAI team to do this)
Reviewing archived variants and learnings regularly to inform future rounds of testing

Autotuning is designed to help you get the best results with minimal manual effort, but it works best when paired with thoughtful test setup and variant approval planning.

What happens if a variant is archived automatically?

If a variant is archived and there are other approved variants available, they will continue running. If no other variants are approved, your control message will be shown instead, which may slow down learning and affect test results.

This fallback ensures the message still sends, but it can slow down learning and impact your test results, since no new variant data is being collected.

To avoid this situation, we recommend:

Approving multiple variants in advance
Monitoring performance thresholds (we also do this for you)
Reviewing archived variants and learnings regularly (we also present this to you weekly)

This ensures the test can continue uninterrupted even if one or more variants are automatically archived due to underperformance. It also helps maintain steady learning and faster optimization.

How are new variants created by Autotuning?

Autotuning automatically generates draft variants by analyzing patterns from your top-performing content. It looks at elements like structure, tone, keywords, and calls to action to create new versions that build on what’s already working. These drafts will appear in your dashboard with a green “Approve” button. Nothing is ever published without your review and approval.

In addition to automated suggestions, you can proactively prompt our AI to create new variants based on what you want to test. For example, you might say:

Create three variants that emphasize urgency
Test a softer tone and a more curious subject line
Try versions that highlight the savings benefit more clearly

This gives you control over what gets tested while still using AI to accelerate ideation and variant creation. You can mix and match your own ideas with system-suggested drafts to get the best of both worlds.

Where do I find Autotuning alerts?

Look for:

A red dot next to the chat icon in Studio
Exclamation marks (!) in your dashboard indicating a test needs review
Slack notifications when action is needed
Updates shared during weekly meetings, or via direct Slack messages if urgent

We’re also working on weekly Slack alerts to streamline this process and ensure you never miss a recommendation.

What is MAB and how does it work in JustAI?

MAB stands for Multi-Armed Bandit, a testing method that dynamically adjusts traffic to higher-performing variants in real time. Instead of splitting traffic evenly across all variants, MAB shifts more traffic to those showing stronger early results.

JustAI supports MAB-style testing, but it is not enabled by default. If you’d like to use this optimization method, contact us and we can turn it on for you. MAB is most effective in high-volume campaigns where early performance signals can reliably guide traffic allocation.

Our system will then use a MAB-like approach for variants that reaches the minimum threshold for performance evaluation (typically 1,000 sends) and:

Reduce traffic to underperforming variants
Increase traffic to top performers
Archive weak variants automatically (if that setting is enabled)

This helps optimize performance continuously, allowing you to maximize engagement while still learning what works.

Using JustAI Chat Assistant

What can I do with the chat assistant in Studio?

You can:

Review flagged variants via the Autotune tile
Ask for alternative suggestions
Pull learnings using the “Learnings” tile
Chat and ask questions about your test.
- Note: The chat currently only has access to the test you’re viewing. It doesn’t reference data or learnings from other tests—this functionality is coming soon.

How do I generate a Learnings report from the chat assistant?

Click the “Learnings” tile and it’ll run this prompt:

“Create a report of the learnings so far. Gather these learnings from the underperforming (archived and active) and overperforming (active) variants. Return a table with: Learnings as copy patterns from the variants, an example variant from the treatment that follows the learning, and a recommendation for the next variant(s) to test.”

This will generate a table that helps you understand performance patterns: what kinds of messaging work well, what doesn’t, and concrete ideas for what to try next.

But that’s just the start. You can chat with the model to dig deeper. Try follow-up prompts like:

What’s the biggest difference between the best and worst performing variants?
Why might this particular variant have underperformed?
Suggest 3 new variants based on what’s worked so far
Which element (tone, structure, CTA) seems to drive performance the most?

You can also ask reflective questions like:

What did you learn from this test overall?
What surprising patterns emerged?
Were there any outliers? Why might they have performed that way?

Treat this like a working session with a strategist. The more questions you ask, the deeper and more tailored your insights will be.

Why is the chat assistant sometimes slow?

The chat response time depends on the size and complexity of your test. Tests with long HTML, many variants, or detailed segmentation take longer to process. We’re working on improving performance and adding clearer indicators.

Analytics and Reporting

Do I need to set up tracking manually?

In most cases, tracking is handled automatically. However, certain ESPs (like SFMC) may require manual placement of open or click tags inside the HTML. If you’re unsure, we can help confirm whether additional setup is needed for your account.

What metrics can I track in the JustAI dashboard?

You can track open rate, click rate, unsubscribe rate, and conversion rate for each test. You can also track custom metrics but those will need to be set up prior to the launch of your test. Contact us to get this set up.

What do the green and red highlights in the dashboard mean?

Green means statistically better performance than the control

Red means statistically worse performance

These signals help you quickly see what’s working or underperforming.

What should I do if I don’t see results or data in the dashboard?

There is a 3-day offset between when your message is sent and when performance data appears in the dashboard. This delay helps ensure we capture complete and reliable engagement metrics before surfacing results.

You can uncheck this if you want to see the incoming data. If you’re still not seeing data, reach out to support and we’ll investigate.

What is the Learnings section and when does it show up?

Once enough data is collected, the dashboard shows segment-specific insights. These include copy patterns, variant examples, and performance lift vs. control.

What is the difference between ‘Active Variants Only’ and ‘All Tested Variants’?

Active Variants Only provides data from the variants currently running
All Tested Variants provides overall data including current and previously approve/archived variants

Troubleshooting and Support

What if I accidentally archive or approve a variant by mistake? Can I undo it?

If you did something by mistake, first check to see if there’s an “Undo” button. If you haven’t saved your changes yet, you can usually undo your last action directly.

If you’ve already saved changes, here’s what to do next:

Accidentally archived a variant?

It’s not deleted: it’s just moved to the archive section. You can unarchive it at any time:
- Find the variant in “Archive”
- Click “Approve” to make it active again.
- Or click “Revert to Draft” if you want to make edits before re-approving.
Accidentally approved a variant?

You can “Reject” it which moves it into Archive, or also “Revert to Draft” to move it to drafts to edit.

Can I test across multiple languages or locales?

Yes. You can test across multiple languages with JustAI. If you’d like help organizing or managing language-specific tests, just let us know. We’re happy to assist with setup and strategy.

Please also let us know which languages you’d like to localize in. While our system can generate localized content, we strongly recommend having a native speaker review and confirm that the translations are accurate and culturally appropriate.

Not seeing data?

By default, there’s a 3-day delay between when your message is sent and when performance data appears in the dashboard. This offset ensures we’re showing complete and reliable engagement metrics.

If you’d prefer to see data as it comes in, you can disable the delay by unchecking the 3-day offset setting.

Still not seeing any data? Reach out to support and we’ll help investigate.

Need help or have a question?

If you’re experiencing an issue or have a question that’s not answered in this guide, we’re here to help.

Whether it’s troubleshooting a test, reviewing results, or planning your next experiment, feel free to get in touch. We’ll make sure you get what you need.