Momentum is one assistant working across your whole media operation. It's built on a controlled set of building blocks — campaigns, creatives, plans, constraints — that combine to handle an enormous range of requests. Every change you approve.
Everything in the matrix below is one of these three. They move from left to right by how much they touch your system — reading, then acting, then removing.
Ask Momentum about any part of the system — campaigns, plans, spots, inventory, history — and it brings back a clear answer. Read is also the foundation for everything else: Momentum can't change, adjust, or remove something it can't first access.
Create new things, update what's there, or adjust how something runs. This is the only category that touches select parts of your system — and only after you say so.
Remove accounts, plans, campaigns, or groups when nothing depends on them. Momentum checks dependencies before it acts and removes from the bottom up.
Key areas of your operation. A filled mark means Momentum works there today. Open any row to see exactly what it can do — in plain terms.
Optimization builds on everything else. If the foundation isn't right, optimization isn't right. It's critical to start with this foundation like we are, to build up usage and continue training before optimization skills start to get introduced.
Every action in the Change and Clean Up columns is proposed first. Momentum shows you exactly what it intends to do, you review it, and only then does it apply. It cannot quietly create, edit, or remove anything on its own.
Momentum is live — a real AI agent connected directly to AWM, ready for your every day work. Early access is a brief period during which we'll focus on quick rounds of AI training based on your usage as well as any feedback shared.
When something comes across your desk — a campaign to set up, logs to check, a question about a show — start by asking or completing it through Mo. The more it's part of your real work, the more useful it becomes.
When a response falls short — it didn't finish the task, answered partially, or left you more confused — that's what we most need to hear. But if Mo handles something in a way that genuinely impresses you, you can flag that too. Both signals go directly into AI training. For misses, the only context we need is what you were hoping to see. We have the rest in the chat.
Every interaction makes Momentum smarter, even when there's no feedback to share. When there is, it goes directly into AI training — so the more your team engages, the better Mo can support the work.
Right inside Momentum — one tap, no separate tool needed.
Wrong action · Incomplete · Couldn't do it · Incomplete information · This was great.
Tell us what you were expecting — or what made it land. We can see the full conversation, so a sentence or two is plenty.
Your input goes directly into AI training. Over time, Mo learns how your team works — and gets sharper because of it.
Most of the ideas your team raised early in planning are covered here — in full or in part. Think of this as a familiar reference point, not a checklist. Some ideas mapped directly; others are covered in ways that go further than the original ask. A few things aren't here yet — some intentionally, some still in progress — and that's exactly what the feedback loop is for.
The fastest way to find what works and what's missing will be using Mo for your actual work and sharing what you notice as you go.
A small number of early ideas weren't included in this release — some require additional integrations, some are in the pipeline, and some may evolve based on how the team uses Mo. These are noted here for transparency.