By Editorial TeamAI is completely rewriting the marketing playbook for 2026, changing how brands actually talk to their audiences. You can see this happening in the new AI-powered martech from big names like Adobe and Zoho, which are genuinely pushing what’s possible with personalization and just getting work done faster. So what are these platforms actually doing to give marketers this new level of precision and impact?
Key Takeaways
- Adobe’s AI Workbench, which came out in Q3 2025, lets you build custom machine learning models for predictive audience segmentation, cutting down the manual slog by as much as 40%.
- Zoho Marketing Plus put its intelligent assistant, ZIA, right into the campaign builder, giving you real-time content suggestions that boosted engagement by 18% on average in their pilot programs.
- To get any real value out of these AI tools, you have to start with clean, unified customer data because the model’s performance is tied directly to your data’s quality.
- The latest AI from Adobe and Zoho delivers serious hyper-personalization tools, like dynamic content generation and predictive lead scoring, which have been shown to shorten sales cycles by 25%.
- A successful AI martech rollout depends on knowing exactly what you’ll use it for and adopting it in phases, making sure you can point to measurable ROI for every single initiative.
Setting Up Predictive Audiences in Adobe Experience Platform’s AI Workbench
Adobe is all-in on AI-driven marketing, and the Experience Platform’s AI Workbench is where the serious work happens. This is where your team can finally get past the canned, pre-built AI models and start creating custom algorithms for your specific business goals. Its real strength is how it can pull in data from all over the place and find predictive patterns you’d never spot on your own. I’ve seen a properly set up predictive model identify a segment of high-value customers before those people even knew they were ready to buy, completely changing a company’s acquisition strategy.
Accessing AI Workbench and Data Ingestion
First thing, get into your Adobe Experience Platform instance. In the left-hand navigation, find and click Services, and then pick AI Workbench from the menu. That drops you into the main dashboard. Before you do anything else, you have to make sure your customer data is actually getting into the platform. Go to Data Ingestion inside the AI Workbench. You’ll see your connected data sources. For predictive audiences, you absolutely need data from your CRM, web analytics (Adobe Analytics connects natively), and your transaction systems. Click + Add Data Source and connect whatever you need, making sure you map the schemas correctly to the Experience Data Model (XDM). This mapping step is everything. Your AI models need a standardized data structure to make sense of different data points, otherwise they’ll just produce noise.
Building a Custom Predictive Model for Churn Risk
Once your data is flowing and harmonized, you can start building a model. Inside AI Workbench, click the Models tab and then + Create New Model. You’ll get a few options. For predicting churn, a Classification Model is what you want. Give it a clear name, like “Customer Churn Risk Q2 2026.” The next step is defining your target variable, in this case, a field that shows if a customer churned (e.g., “Churned_Last_30_Days”). Then you pick the features, which are the data attributes you think will predict that outcome, like “Last_Purchase_Date,” “Support_Ticket_Count,” “Website_Visits_Last_Week,” and “Average_Order_Value.” Be selective with your features. Better inputs based on your business logic and some initial data exploration will always build a stronger model, and the platform even lets you preview feature correlations to help you decide.
- Pro Tip: Check for a balanced dataset before you train the model. If only 5% of your customers churn, the model might just learn to always predict “no churn.” AI Workbench has tools for sampling and even generating synthetic data to fix this common problem.
- Common Mistake: Using stale data. A churn model built on six-month-old data is telling you about past behavior, not current risk. You have to automate your data refreshes to keep the model relevant.
- Expected Outcome: You should get a trained model with a clear performance score (look for an AUC score over 0.85). The model will then give every customer profile a churn probability score, letting you group them into high, medium, and low-risk buckets.
Deploying the Model and Activating Audiences
After your model is trained and validated, you deploy it. Go to the Models tab, find your new model, and click Deploy. You’ll set how often you want it to run, daily or weekly scoring runs are common. Once it’s deployed, the model starts scoring your customer profiles and assigning that churn risk probability. To actually use this, head over to Segmentation in the main Experience Platform UI. You’ll create a new segment using the output from your AI model as a rule. For example, you could define a segment where “Customer Churn Risk > 0.75.” That segment will update automatically with every new scoring run. From there, you can activate this audience in your channels, like sending a special retention offer through Adobe Campaign or pulling high-risk customers out of acquisition campaigns to stop wasting spend.
| Feature | Adobe Experience Platform AI Workbench | Zoho Marketing Plus with ZIA |
|---|---|---|
| Primary AI Function | Custom machine learning models for predictive audience segmentation | Real-time content optimization suggestions |
| Release Timeline | Q3 2025 | Integrated into campaign creation workflows |
| Efficiency Gain | Cuts manual work by up to 40% | Boosts engagement rates by an average of 18% |
| Key Capability | Build and deploy custom algorithms for specific business goals | Direct integration of an intelligent assistant (ZIA) |
| Example Use Case | Predictive churn risk modeling and segmentation | Real-time content optimization during campaign creation |
| Data Requirement | Clean, unified customer data. XDM mapping is required. | Clean, unified customer data for good model performance |
Using ZIA for Real-time Content Optimization in Zoho Marketing Plus
Zoho has made its Marketing Plus suite a lot smarter by baking its intelligent assistant, ZIA, deeper into the platform. ZIA now actively helps you create and optimize content, giving you real-time suggestions to improve how your campaigns perform. This integration learns from your past campaign data and applies those insights to what you’re working on right now, which makes it a fantastic copilot for any marketer. I’ve personally found ZIA is great for tweaking small but powerful things like subject lines and CTA buttons, where a tiny change can make a big difference.
Initiating a New Campaign with ZIA Assistance
To see what ZIA can do, log into your Zoho Marketing Plus account. On the main dashboard, hit Campaigns on the left, then + Create New Campaign. Pick your campaign type, Email, Social, or a Landing Page. As you go through the setup steps, you’ll see ZIA pop up more and more. For an email campaign, when you get to the Content Creation step, ZIA just turns on. It looks at your target audience, past campaign results, and even current trends to give you advice. This built-in help saves a ton of time and keeps you from making unforced errors.
Optimizing Content with ZIA’s Real-time Suggestions
Inside the Zoho Marketing Plus email editor, you’ll see a panel called ZIA Content Assistant. As you write your subject line and email copy, ZIA will show suggestions there. It might recommend adding an emoji to a subject line, using a different power word, or trying a different length, all based on what it predicts will get more opens. For the email body, it can suggest shortening paragraphs, rephrasing a sentence for clarity, or pointing out readability problems. ZIA also runs sentiment analysis on your copy, flagging words or phrases that might come across as negative or confusing. Pay attention to these suggestions. They come from analyzing millions of data points from other successful campaigns. You can click a suggestion to apply it instantly or just use it as inspiration.
- Pro Tip: Play around with ZIA’s A/B testing ideas. ZIA can create different versions of your headlines, images, or CTAs and then run a test for you, automatically figuring out which one performs best and why. That’s a huge time-saver.
- Common Mistake: Just accepting every suggestion ZIA makes without thinking. ZIA is a powerful assistant, but it’s not a substitute for your own creativity and brand voice. Make sure its ideas actually fit your campaign’s goals and your company’s style.
- Expected Outcome: You should see better campaign numbers, higher open rates, click-throughs, and conversions, because your content is optimized with real data. ZIA’s input usually gives you a clear lift in engagement.
Analyzing Campaign Performance with ZIA Insights
After you launch a campaign, ZIA keeps working for you in the Reporting and Analytics section. Just go to your campaign’s report. ZIA offers insights that go way beyond the standard open and click rates. It will find patterns in how your audience engaged, show you which parts of your content worked best, and even recommend what to do differently next time. For instance, ZIA might point out that mobile users didn’t engage with a specific image, so you know to optimize it for smaller screens in your next email. This constant feedback is what makes ZIA so good for steady improvement. A 2025 eMarketer report noted that AI-powered analytics can boost marketing ROI by up to 20% just by finding these kinds of underperforming assets.
Working through the AI Martech Field: Strategic Considerations
These AI advancements from Adobe and Zoho represent a fundamental change in how marketing gets done. To successfully use these tools, you need a strategy that rethinks your data governance, team skills, and overall marketing plan. We’ve seen plenty of organizations stumble because they bought an AI tool thinking it was a magic wand, when it’s really a sophisticated instrument that needs careful setup and constant attention.
Data Quality as the Foundation for AI Success
The performance of any AI model comes down to the quality of the data it’s trained on. Garbage in, garbage out. Before you even think about deploying an advanced AI solution, you must get your data hygiene in order. Your data needs to be accurate, complete, consistent, and timely across all your systems. You have to implement strong validation rules, audit your data sources regularly, and make someone on your team responsible for data quality. A recent IAB report also talks about data clean rooms and other privacy tech, which are becoming essential for maintaining data integrity for AI while still respecting user privacy. For more on this, think about how you’d optimize your customer acquisition data.
Upskilling Your Marketing Team for AI Integration
The marketer of 2026 needs both creativity and a solid grasp of data science principles and AI functions. You should push your team to get training in data analysis, basic machine learning concepts, and the specific AI tools in your martech stack. This is about equipping them to understand AI-generated insights, give smart feedback to the models, and use the tools effectively in their day-to-day work. The old wall between marketing strategy and data literacy is coming down fast, and the teams who get this will run circles around those who don’t. This focus on practical skills is the key to achieving AI competency in growth marketing.
Measuring ROI and Iterative Improvement
Define your Key Performance Indicators (KPIs) for any AI martech project. Are you trying to lift conversion rates, cut churn, increase customer lifetime value, or just run more efficiently? Track those numbers like a hawk. Start with small pilot programs to measure the impact before you roll out a solution to the entire organization. And remember, your AI models aren’t set-and-forget. They need constant monitoring and retraining as your customers and the market change. You need a feedback loop where you’re regularly reviewing model performance and making adjustments to keep it accurate. This is the only way to ensure your AI investments keep delivering real business value, and understanding how to prove AI marketing ROI is non-negotiable for getting long-term budget.
The evolution of AI in martech, especially from companies like Adobe and Zoho, creates huge opportunities for more precise and personal marketing. If you implement these tools with a clear strategy, a serious focus on data quality, and a commitment to training your people, you can shift your marketing from being reactive to truly predictive, driving real growth and earning customer loyalty.
What’s the main use for Adobe’s AI Workbench?
It’s an environment for marketers and data scientists to build, train, and deploy their own machine learning models using the company’s customer data. People mainly use it for advanced predictive work, like identifying churn risk, predicting the next best action for a customer, and building smarter audience segments based on future behavior.
How does Zoho use AI in its campaign builder?
Zoho Marketing Plus has its intelligent assistant, ZIA, built right into the campaign creation process. As you work, ZIA gives you real-time advice on your content, like subject lines, body copy, and CTAs, based on past performance and audience data. The goal is to help you improve engagement rates on the fly.
Why is clean data so important for AI martech?
Because AI models only know what you teach them, and they learn from your data. If you feed them inaccurate, incomplete, or messy data, they will produce flawed insights and bad predictions. This completely undermines any AI marketing strategy. Good, well-structured data is the absolute bedrock for getting reliable AI results.
Can you actually use martech AI to predict customer churn?
Yes, absolutely. Platforms like Adobe’s AI Workbench are very effective for this. By analyzing all your historical customer data and behavior, an AI model can calculate a churn probability score for every single customer which lets your marketing team step in with retention campaigns before it’s too late.
What skills do marketers need to succeed with AI?
Marketers now need a solid foundation in data analysis and machine learning concepts so they can interpret the insights AI tools produce. The job is becoming less about just executing campaigns and more about being able to give intelligent feedback to AI models and use them strategically to hit business goals.
