In this article
Marketing teams are getting pulled in two directions at once. The pace keeps climbing, with more campaigns, more channels, and more tests on the list, while the work that makes a message land still takes the time it takes: your voice, and your judgment about what's worth sending. AI has shown up to help, but a lot of it arrives as one more place to check, a copilot in the editor here and an assistant on a dashboard there.
Our fall 2026 release brings everything together, helping you pull off even smarter marketing. Your AI Agent can now answer questions about your workspace data, suggest experiments for your automations, and draft in Design Studio, which is now the default email editor. Each one feeds the next, so what you find in your metrics shapes the test you run, and the result of that test shapes the next thing you write.
Ask your Agent what to test next, and it suggests specific experiments for your automations. That frees up more of your time for the decisions only you can make, like which test is worth running and whether the copy still sounds like you.
Here's what's new, how to use each update today, and prompts you can paste straight into your workspace to get started.
Pro tip
Join us live on Thursday, November 19 at 11:00 am ET for a walkthrough of everything in this release. Register here.
TLDR
- Your AI Agent can now answer questions about your workspace data, so asking what's behind last week's dip in open rates returns the numbers and the context in one answer.
- Experiments in Customer.io now run across branches within an automation, so you can compare paths such as email-only versus email plus SMS, add a holdout to see lift over doing nothing, and keep every result on record.
- Design Studio is the default editor for email and your Agent can draft the first version from a plain-language description. Build for notification inbox, SMS, and WhatsApp in Design Studio too, with more channels coming soon.
- Seven more updates are part of our fall 2026 release, including sign-up forms, consent records, a native Databricks source, and per-channel rate limits for automations.
Agent analytics: Ask what's driving changes in your metrics

When open rates dip, the usual routine is to open the dashboard, slice by automation and date, export a CSV, and start squinting.
What's new in this release
- Your AI Agent now answers questions about your workspace data, so you can ask what's behind last week's dip in open rates and get the numbers and the context in one answer.
- It queries live data on deliverability, site activity, pipeline health, and profile changes across every channel you send on, from email and SMS to push.
Here are six prompts to test out Agent analytics yourself:
Diagnose a dip
“Open rates on our weekly newsletter fell last week. Which sends account for the drop, and is it concentrated in one inbox provider like Gmail?”
Find what's working
"Which automations drove the most conversions last month?"
Catch disengagement early
“Which of our automations saw the biggest drop in click rate over the past 60 days?”
Check delivery health
"Has our email deliverability changed over the past two weeks, and what's behind it?"
Attribute SMS
"How many purchases did our SMS messages drive in the last 30 days?"
Pick a test audience
“Which automation has the most volume but the lowest conversion rate? I want to test there next.”
Name the automation and the date range in your question, as in "the welcome automation, last week," and follow up with "why?" when something looks off. Finish with "what should we test next?" to hand the finding straight to experiments.
Important to note: purchase attribution looks back 90 days and is last-touch within 30 days. SMS reports don’t include opens.
Experiments: Test whole automations, then ask the Agent what to test next

Testing is usually the first thing to slip. Picking a test, setting it up, tracking it, writing up the result, and choosing the next one adds up to a full-time discipline that most of us run part-time, so when the week gets busy, the idea goes back into the backlog. A single-message A/B test also tops out at telling you which subject line won, and the questions that move lifecycle numbers tend to be bigger. Does adding SMS improve conversion? What changes when you wait a day between sends?
What's new in this release
- Experiments in Customer.io now run across branches within an automation, so you can test whole paths, from one email to a full sequence.
- You can add a holdout to see your lift over doing nothing.
- You choose the success metric for each test, such as opens, clicks, conversions, or retention, and it doesn't have to match the automation's goal. A welcome automation can be judged on activation while a winback automation can be measured on whether or not people come back within 60 days.
- Every result stays on record, so you can build on the last test.
- Ask your Agent what to test next, and it suggests specific experiments for your automations.
Before you build any experiment, define:
- The one question the test should answer
- The metric that reflects the outcome you care about
- Whether to add a holdout
Then bring in the Agent for suggestions, and you choose which experiments to run.
Here are some prompts you can test to get started:
"What should I test next in my welcome automation? I care about activation."
Focus on an outcome
"Look at my cart abandonment automation and suggest an experiment. Measure it on completed purchases."
Test a channel
"Does adding an SMS two days after the second email improve conversion in my trial expiry automation? Set up that experiment with a holdout."
Test timing
"We wait three days between the first and second onboarding emails. Suggest a test on that gap, measured on activation."
Keep the cycle going
"The test on my winback automation has finished. What should we try next?"
Design Studio: Your email editor, with agentic content creation

Design Studio is the editor in Customer.io where you create beautiful, on-brand emails at scale without the complexity. It combines intuitive visual tools with developer features, so your team has everything it needs to build high-quality messages in one place.
Design Studio is now the default editor in Customer.io. It's where you think, draft, review, and ship, with agentic drafting built in, inbox previews that catch issues before you send, and your voice intact throughout. Agentic drafting cuts the time to a finished email, and the review step keeps your voice, so your Agent is the crew, and you stay the director.
What's new in this release
- Design Studio is now the default email editor in Customer.io. Build for notification inbox, SMS, and WhatsApp in Design Studio too, with more channels coming soon.
- You describe what you want, your AI Agent drafts it in your brand, and you make the final call.
- You can ask the Agent for changes from inside the editor canvas.
- Inbox previews show how the message will look before you hit send.
Give the Agent what you'd give a good freelancer: the audience, the goal, the one action you want the reader to take, and a length limit. Here are five prompts to start with.
Write a first draft
"Draft a welcome email for people who started a trial today. Friendly tone, one button that sends them to connect their first data source, and our brand colors."
Rescue an abandoned browse
"Write a browse abandonment email for the product someone viewed but didn't buy. Keep the body under 80 words and put the button above the fold."
Revise in place
"Shorten the headline to six words and make the second paragraph sound less formal."
If you have SMS or WhatsApp connected, try drafting this
"Write an SMS reminding someone their cart is waiting. Stay under 160 characters and end with a friendly sign-off."
Draft a WhatsApp message
"Write a WhatsApp message confirming that an order has shipped and sharing the tracking link."
Treat every draft as a first draft. Always read the copy against your brand voice and check the inbox preview.
Also in this release
Sign-up forms
Capture email addresses and phone numbers and turn visitors into leads. Forms now take flexible inputs, including open text fields, and come with conversion-focused analytics so you can see which ones pull their weight.
Consent records
Keep a record of when, where, and how each person gave consent. If you send into regions with strict consent rules, that takes a lot of the stress out of audits.
AI-native integration setup
Click "Set up with AI" and your AI Agent connects your mobile app, web app, or server to Customer.io. Browse the full list of integrations to see what you can connect.
Databricks source integration
If your data lives in Databricks, you can now connect it directly to Customer.io as a native reverse ETL source.
Automation rate limits
Add a per-channel limit to any automation or API-triggered broadcast to protect your sender reputation or keep from flooding downstream systems.
Object Trigger
Start an automation when a person, their relationship, and a related object all match your conditions, in any order.
Customer.io certification
Earn your Certified Practitioner certificate to show you can run lifecycle marketing in Customer.io, from clean data to triggered automations to healthy deliverability.

What to do next
These updates are available on all plans Essentials and up, so you can try them in your workspace today. Start with the one closest to a gap you've been working around, and if you're not sure where to begin, ask your Agent what to test next in your highest-volume automation.
Join us live on Thursday, November 19 at 11:00 am ET for a walkthrough of everything in this release. Register for the webinar.
Read all about what the Agent can do, or book a demo for a walkthrough of everything you can do with Customer.io.
FAQ
Can the Agent draft emails for me?
Yes, Design Studio includes Agent drafting. You describe the email, the Agent proposes a first version in your brand, and you review and refine it in the editor, with inbox previews to check how it will look before you send.
Can I test an entire automation, or only individual messages?
You can test entire branches within an automation, from one email to a full sequence. You set the success metric, such as opens, clicks, conversions, or retention, and you can add a holdout to measure lift over doing nothing.
What can I ask the Agent about my data?
Anything in your workspace, including deliverability, site activity, pipeline health, profile changes, and how messages drive purchase events across channels. Purchase attribution looks back 90 days and is last-touch within 30 days. SMS reports don’t include opens.
Does the Agent run experiments on its own?
Experiments run only when you decide to launch them. You ask the Agent to suggest experiments for your automations, then you choose which ones to run and which metric each one is measured on.
What happens to the classic editors?
New accounts created from November 5 get Design Studio only. Existing accounts keep access to Customer.io's classic email editors.
Are these updates available on every plan?
Yes, these features are available across all Customer.io plans Essentials and up.
Drive engagement with every message
- Omnichannel campaigns
- Behavior-based targeting








