AI Growth Marketer
AI Growth Marketer
The Maxma AI growth marketer is an agent that runs marketing work end to end across your ad platforms, your store and your analytics: finding what is working, making the creative, building and launching the campaigns, fixing the tracking underneath, and reporting on what changed — continuously, with you deciding what goes live.
Why it is built this way
One picture, not a dozen tools. Marketing work is spread across ad platforms, a store, analytics and a tag manager, and most AI tools see one of them at a time. Maxma has the whole picture and the ability to act on all of it, so it can reason about what is actually happening and carry the work through, instead of handing you a suggestion that stops at the edge of one tool.
You keep control, not the manual work. The hours go into the low-leverage, time-consuming parts: pulling the numbers, building the campaigns, resizing the creative, checking the tags. Those are the parts you hand over. The decisions stay yours, so a small team gets the output of a much larger one without giving up the calls that matter.
It gets on with it. Work does not depend on you remembering to ask. Maxma connects to the major ad platforms, commerce and analytics tools, with more added regularly, and it works over weeks rather than in single sessions, carrying what it learned from one round into the next. The craft it applies is encoded from people who have run this work at scale.
The building blocks
The work itself breaks into a set of building blocks: analysing the data, making the creative, building and changing campaigns, fixing the tracking, writing for the store, delivering the result. They mix and match into almost any piece of paid marketing work. Ask for one on its own, string several into a workflow, or let a co-pilot reach for them without being asked.
| Building block | What it covers |
|---|---|
| Analysis | A written analysis with charts and tables, pulling together every connected source — ad platforms, analytics, Search Console, your store — around the question at hand |
| Dashboards | A dashboard built around the metrics that matter, across the same connected sources, laid out the way you want to read it |
| Creative generation | A UGC-style video, a product shot, a static ad — made from the assets you already have in your store, your site or an upload, and ready to run as an ad, list in your store or sit on your page. Built on the leading image and video models, with our own editing layer on top |
| Creative tagging | Ads tagged on whatever dimensions matter — hook, format, talent, offer, product — with performance compared across each |
| Competitor research | What competitors are running on Meta and TikTok, and any of those ads recreated for your brand |
| Keyword research | Volumes, competition and costs, grouped into ad groups worth buying |
| New campaigns | A campaign on any of 8+ ad platforms, in the types that platform supports: search, shopping, performance max, social, video and more, down to the ad groups, keywords, audiences and ads |
| Campaign changes | Add an ad group, an ad or a keyword; rewrite copy; adjust budgets, bids and bidding strategy; tighten targeting, audiences or geography; add negatives; pause what is not working or turn something back on |
| Bulk ad upload | A folder of assets turned into ads, matched to products and copy |
| Tracking setup | The exact tag, pixel, conversion or key event change, in your tag manager or analytics |
| Store listings | Products, collections and pages written for your store, from your own photos and details |
| Landing pages | A page built to match the ad that brought the click |
| Delivery | The finished work posted to Slack or sent by email, where your team already is |
The full list is on capabilities. What the agent may put live by itself, and what waits for you, is yours to set: see write permissions.
Three ways to use it
Ask for one thing. Describe what you want in a message and get the finished work back. See on demand.
Write down work that repeats. A workflow runs on a schedule or when something changes, and delivers the result where your team works. See workflows.
Hand over an area. A co-pilot takes a scope and a brief and works on it continuously, bringing you proposals rather than questions. See co-pilots.
Who decides what goes live
Anything that only produces work inside Maxma — reading your accounts, analysing data, writing a report, drafting creative — happens without asking, because nothing outside Maxma changes.
Anything that touches a live account is governed by permissions you set, class by class: budgets and bids, campaign structure, creative going live, targeting, keywords, tracking, store and catalog. Set a class to wait for you and nothing in it reaches an account without a person. Set it to run and the agent carries it out and reports.
Video is past its saturation point while TikTok prospecting sits on the steep part of its response curve. This shifts the budget across five campaigns — net weekly spend unchanged.
Video_Prospecting_NAcampaign-25%- Video_Prospecting_EUcampaign-25%
- Video_Retargeting_NAcampaign-50%
Prospecting_Broad_NAcampaign+40%- Prospecting_Interest_EUcampaign+40%
- Net weekly budget change$0
The settings are yours, and the agent cannot work around them. See write permissions.
How it knows what is working
By default the agent judges performance the way the ad platforms report it, which flatters the channels closest to the click. Connect Google Analytics and it has a second view. Run incrementality tests and it has measured lift for the channels you tested. Add the marketing mix model and it has an ongoing read on incremental performance across the whole portfolio, including the channels no test can hold out.
Those sources rarely agree. The agent can reconcile them: which number to trust for the decision in front of you, and what it means for where the next dollar goes. The more measurement you connect, the closer its target gets to real growth rather than what is easiest to track.