The daily account work, with the manual parts taken out.
NYX tells you where spend should move and shows the working, generates the creative to test next, and answers the numbers questions people keep sending you. You approve everything. Nothing changes on its own.
Connect a live account and check the recommendations against your own read.
Connects to the accounts you are already in all day.
Campaigns run across Google, Meta, LinkedIn, TikTok and Snapchat. Reporting connects your ad accounts to Shopify, GA4 or your MMP.
You already know what a good account looks like. The problem is that doing the work properly, across every campaign and every platform, every week, competes with actually building and launching things.
- The optimisation pass slips because building next week's campaigns is louder
- Creative testing is capped by how fast someone else can produce the next variation
- Half the day goes to pulling numbers for someone who asked a question you have answered before
- Most automation tools want to change your account for you, which is exactly what you do not want
What NYX does for you.
Recommendation Engine
A prioritised list every morning, built from your own account's last three months, with the calculation behind every call visible.
Pixeo
Brief a variation in chat, review the direction it proposes, approve, and get a launch ready creative back. Each version is saved separately, so a test round is a set of versions you can track, not a folder of overwrites.
Neo
Ask what moved and get the answer with the chart. A CAC spike traced back to CPC, CTR or conversion rate without opening four tabs.
Xeno
Builds and launches across five platforms from the same workspace, so a cross platform test is one build session instead of five.
Custom Reports and Automated Alerts
Save the answers people keep asking for, and let the alert reach them instead of you.
How it works.
Connect
Link Google and Meta, plus Shopify, GA4 or the MMP so performance can be read past the click.
Set brand context
Brand details and guardrails go in once on the brand onboarding page, and creative works off them from then on. Editable whenever the brand moves.
Work the recommendation queue
Daily prioritised list with the reasoning shown. You accept, skip or defer.
Test and measure
Generate the next creative round in Pixeo, launch through Xeno, read the result in Neo, start the next round from the winner.
Inside the Recommendation Engine.
The detail that matters, because you will judge this on whether the recommendations are any good.
- Compiles an SOP for the brand from the last three months of its own data, then regenerates a prioritised list daily. Recommendations reflect how that account behaves, not a generic best practice sheet.
- Runs at account, campaign, ad group and ad level, across budget, bidding, creative and audience.
- Every recommendation shows the expected impact, how it was calculated, and the 14 day trend of the metrics it touches. You can disagree with it on the evidence rather than on faith.
- Learning phase holds, data quality gates and maturity thresholds are built into the rules, so nothing fires on a campaign with too little data to read.
- Deterministic. The language model phrases the output. It does not generate or alter a number, a score or a recommendation.
- Nothing is applied automatically. Ever.
What changes.
The optimisation pass stops slipping
The queue is built before you sit down, so the work starts at deciding rather than at looking.
Creative rounds stop waiting on someone else
Brief the variation, approve the direction, launch it the same day.
The question queue gets shorter
People pull their own numbers out of Neo instead of pinging you.
The recommendation logic on this page is the same one the NYX agency team runs on live client accounts across D2C, streaming, education and app businesses.
Platform, agency, or both.
If you are running accounts solo and need cover, the NYX team can take a channel or a client and work inside the same workspace. If you are in house, the platform is yours to run. Either way, approval stays with you.
Questions about NYX for Performance Marketers
Will NYX change my account automatically?
Where do the recommendations come from?
Do recommendations account for learning phase and thin data?
Is Google and Meta performance merged into one number?
Can Pixeo produce a batch of variations at once?
What can Neo reliably answer?
Need more help?
Our team is here to answer your questions.
Connect an account you know well and check the recommendations against your own read.
Connect a live account and check the recommendations against your own read.