The Forecaster Summer 2026 is here

In this edition

• Prescient AI: Your new AI assistant, built into the platform.
• Media Forecaster: One plan across every model, with the reasoning behind each recommendation.
• Model Center: One hub for every model you run, including health, signals, and accuracy.
• Data Center: Self-service controls for your integrations, connectors, and campaign mapping.
• Attribution: Base and Halo revenue side by side, next to platform-reported numbers.
• Onboarding: A guided setup that shows only the connectors your account needs.

The Forecaster Summer 2026 is our biggest release of the year: six updates built around one idea. Each update gives you more direct control over your model, budget, and data, still rooted in the idea that all of your marketing efforts have a base performance and halo effects. Everything below is live in your account today.

Get ready for less time staring at dashboards and more time focused on getting results.

AI ACTION CENTER

Prescient AI

Understanding what your model is telling you and knowing what to do next has usually meant a call with your CSM or an afternoon poring over the dashboards. Prescient AI changes that.

Meet your new AI assistant, now built into Prescient AI. Ask a question about your own numbers in plain language and get an answer grounded in your brand's unique model. Recap the last 30 days, ask why a channel's performance moved, or compare this month to last. Prescient AI reads what your model already shows and explains it back to you.

Prescient AI works alongside your CSM, who will remain your partner for the strategic calls. Every answer stays inside your own data and traces back to your model, so Prescient AI never makes up a number or uses industry benchmarks.

PLANNING

Media Forecaster

Planning used to mean two tools and two answers. We've reconciled that system so you get one recommendation based on your brand's historical data, making it easier than ever to present Finance with a plan.

Media Forecaster optimizes and forecasts at the same time. Set a goal (whether it's revenue, ROAS, new customers, or cost per new customer), and get a single plan across every model you run. You'll see the reasoning behind each recommendation before you commit.

Now you can walk into the budget meeting with defensible next steps, built on the model you already trust.

TRANSPARENCY

Model Center

Trusting a model you can't see into is hard. In the past, you often needed CS or an analyst to explain what was happening with the model when a recommendation changed.

Model Center puts that understanding in front of you: one hub for every model you run, including its health, the business signals feeding it, its full configuration, and a model accuracy score from 0 to 100 that shows how closely it tracks your real sales. When a number moves, you'll see what changed and why in time for your next meeting without waiting on a ticket.

SELF-SERVICE

Data Center

Every marketer knows the drill: you launch a new campaign, then wait on a ticket or an engineer before your model reflects it.

Data Center hands you the controls. Now you have one self-service screen for the setup behind your model and your integrations, connectors, and campaign mapping. When something changes, you map it yourself in minutes and keep moving. Customer success stays available for the complex work, while the routine changes stop requiring any wait time.

Fix your own mapping, on your own schedule.

MEASUREMENT

Attribution

Your platform dashboard reports one ROAS. Your model reports another. You used to be the one left reconciling the two by hand, weighing out whether the channel that looked weak on the dashboard was lifting everything around it.

Now Attribution puts the full picture on one page. For every channel you see the revenue it drove directly and the lift it created elsewhere (Base and Halo side by side), next to the platform-reported number. Sort it, filter it, and export it to CSV to share with the team during your planning meeting.

Modeled ROAS and platform-reported ROAS answer different questions. Attribution shows you exactly where they diverge so you can put budget behind what's working.

GETTING STARTED

Onboarding

Getting started used to mean staring at a catalog of connectors, guessing which ones you needed, and handing the rest to a developer with no clear view of what was done or what was a blocker.

The new onboarding is guided. It shows only the connectors your account actually needs, tracks your progress to a clear finish, and flags issues early, so a missing permission or a misconfigured source surfaces before it can hold up your first model. You always know what's done and what's remaining.

Every release is included in your current plan, at no additional charge.

Over the next few weeks, we'll send a closer look at each release.

See you in the next one,

The Prescient AI team

P.S. Want a walkthrough tuned to your account? Ask your CSM and we'll set it up around your model.