SensibleAI Studio

We are making SensibleAI faster to build with, enabling deeper integrations, and simplifying implementations. And it all starts with documentation. dev.onestream.com is the central launchpad for developers and implementers building with SensibleAI.

This is your portal for discovering what's available, understanding how the pieces fit together, and building with confidence. Written by developers for developers, by implementers for implementers. Because SensibleAI is not just something you configure, it's something you can build with, extend, automate, and integrate.

And we're showing you how. Now, with that foundation, what are you actually using? That brings us to SensibleAI studio. Today, studio contains more than 60 routines, and that catalog continues to grow.

These are the reusable building blocks that bring AI capabilities directly into your OneStream solutions. What matters just as much as the routines is the ecosystem forming around them. There's documentation to understand what exists, coding assistance paired with Developer Studio to accelerate how you build, and a growing set of tools designed to take you from idea to production faster.

All of this working together to make building production-grade AI capabilities simpler and the insights they deliver more powerful. Let's take this forward to SensibleAI forecasts. Today, developers can already programmatically control forecasts in C#.

For many organizations, it's not just a solution that you configure and interact with in the UI, but it's a capability you orchestrate, automate, and embed directly in your workflows. And now, we're taking that builder experience even further.

We're expanding how developers and power users can interact with sensible AI forecasts via typed clients. So from any service that can make a REST call, you can interact with the capabilities and intelligence of sensible AI forecasts.

Whether you're thinking of your analytics platforms, in-house solutions, or your broader enterprise software suite, Forecast fits right in. Putting all these pieces together, the picture is quite powerful.

We've been hard at work, not just creating better AI capabilities, but better ways to build with them. Staying on the forecast topic, we know that not every forecast challenge is a developer challenge.

For organizations, the challenge is often different. How quickly can we go live? How much setup is required? How much work stands between today and my first useful forecast? Well, that gap between today and a first useful forecast is exactly what we went after.

So I'm extremely excited to be introducing SensibleAI Forecast Express. A simpler, faster path to value with SensibleAI Forecast. Now, today, getting live on sensible AI forecast is a meaningful implementation project.

Configuration can be complex, take time. And for many organizations, especially those earlier in their AI journey, that can be enough to pump the brakes. Forecast express rewrites that story. Not a separate product, but a newly supported implementation methodology.

The unlock is a native cube integration. Organizations can now source historical data directly from the cube and write forecasts right back to it. No custom data pipelines, no extra maintenance. Let's see it in action.

SensibleAI Forecast express. Let's assume I am a power user at Golfstream, a vertically integrated golf retailer. I'm preparing to kick off my revenue forecast for drivers in woods for our West Group entities.

It's January, 2026, and I've just loaded my actuals for December, 2025. Sensible AI forecast allows me to source data from and write forecast results back to my cube without needing to write custom data pipelines.

Within my newly created project, I can configure my source data. Forecast express allows me to select my cube and scenario and the time periods I want to pull my data from. Since more data generally means the better model performance, I'll pull in seven years of history.

I will select the rest of the members to pull data from, filtering results by pinning specific members and selecting the actual line items I want to forecast for. With my dimensions selected, I can generate a preview of my data to ensure it's the data I want.

I can see the number of line items I'm forecasting for and the amount of historical data I have for them. Once confirmed, I can submit. The new cube administration page allows me to configure the workflow responsible for loading my forecast results.

The workflow page displays the various workflows, transformation rules, and data sources in my application. Let's create some new ones for my project. I can see the existing data sources filtered by cube and scenario type.

I'll create a new one, providing a name, sensible AI forecast, and selecting my thin detail cube and forecast scenario type. Next, I will create a new transformation rule profile. Finally, for the workflow profile, I will provide a name, select my cube root workflow profile, and submit.

Any new sources will be created automatically based on my configurations, and my workflow profile is now ready to load my forecasting results. A streamlined output configuration allows SensibleAI Forecasts to write the results back to the cube as the prediction completes.

Now let's switch perspectives. As an end user, cube views are my home base, and my SensibleAI Forecast results are already waiting for me here. It's never been easier to access my generated forecast insights.

Looking out a few months, let's drill down into the Phoenix Mach 10 forecast for April. Drilling back, I can load the prediction summary view from sensible AI forecast. With the full year's forecast in view, I'll hone in on April.

Below, the prediction explanation for my April forecast shows how the features impacted this result. I can view the tug-of-war chart to see the feature contributions across my forecast horizon, the features that increased or decreased my forecasted value.

The periodic explanations page shows an alternative view. I'll select a forecast start date and forecast name. For each feature group and the comprising features, I can see the contributions made by each across my forecast horizon.

That is Forecast Express. From historical data in the cube to a machine learning forecast written back. Streamlines, native, no custom data pipelines. Zooming back out, Forecast Express slots right into the broader sense play of the story we started with.

Building faster, integrating deeper, and implementing easier. Further augmenting the platform you deserve.