FA
SYS_ONLINE · Berlin / Remote · Open to advisory + senior IC roles

Fayez Abou Hassoun

Agentic AI for Performance Marketing

I architect autonomous decision systems that ingest noisy ad data, diagnose what's working, and ship spend decisions across Google, Meta, TikTok, Snapchat & Bigo — at the scale of $3.5M+ across 30+ countries.

See the AI OS Browse projects
$0
Managed Spend
0
Countries Active
0
In Performance
∞
AI-Powered Systems
01Identity

The operator behind the agents

FA
Fayez Abou Hassoun
Head of Performance Marketing
MobileArts  ·  Berlin
Online

Performance marketer who treats every campaign as a control system. For seven years I've accelerated growth across mobile, fintech, MVAS & DCB — and for the last two I've been replacing dashboards with autonomous agents that read the data, decide the move, and learn from the outcome. Less reporting, more compounding.

Google Ads Meta Ads TikTok Programmatic MVAS / DCB AI Agents Automation Leadership
Lat / Lon
52.5200° N  ·  13.4050° E
Stack
LLM + SQL + Ads APIs
Ops since
2018

Operating principles

  1. 01
    Spend is a control variable, not a metric.
    Every dollar gets a thesis, a hypothesis, and an exit condition.
  2. 02
    Ship the loop before the dashboard.
    If a human has to read it to act, it's already too slow. Agents close the loop.
  3. 03
    Approval > Autonomy.
    Agents decide. Operators approve. Risk stays bounded, learning stays fast.
  4. 04
    Cross-channel, single brain.
    Google, Meta, TikTok, Snap, Bigo — one decision graph, one performance memory.
02The Method

An AI Operating System for performance

Five-stage agentic loop that turns raw ad-platform exhaust into bounded, reversible, learning decisions.

01

Ingest

Pull cost, conversion, creative, geo & carrier signals from every ad API + first-party event stream into one normalized graph.

Ads APIsBigQueryS2S
02

Diagnose

LLM agents read the graph in plain English: what slipped, why, where the leak is, which segment is compounding.

LLMCohortingAnomaly
03

Decide

The Key-Action model picks bid, budget, creative & geo moves with confidence intervals + expected lift attached.

Key-ActionBanditsHeuristics
04

Approve

Operator reviews proposed moves with full reasoning trace. One tap pushes changes live across every connected channel.

SlackTelegramAudit
05

Learn

Outcomes feed back into the decision graph. Wins compound. Losses become guardrails. The system gets sharper every cycle.

MemoryEvalDrift
03Selected Work

Systems shipped, not slides.

Each project is a real production system or program — currently running or recently completed.

P/01
In Beta
SaaS

Google Ads Automation SaaS

Cross-campaign-type optimization layer over Google Ads. Search, PMax, Demand Gen & Video share one budget brain — bid floors, negative-keyword propagation & creative rotation handled by agents.

Google Ads API SaaS Cross-campaign
4campaign types
−27%wasted spend
1-clickactivation
P/02
Shipping
Flutter

Cross-Platform Learning App

AI-powered learning app built in Flutter. Adaptive curriculum picks the next lesson per learner; LLM tutor handles explanations, hints & assessments offline-first.

Flutter LLM Tutor Adaptive
iOS+ Android
Offlinefirst
Adaptivepath
P/03
Internal
Framework

Key-Action Model System

Decision framework that maps every campaign signal to a small, bounded set of "key actions" the agent is allowed to take — auditable, reversible, and easy to reason about.

Framework Decision graph Guardrails
14key actions
100%reversible
Auditby default
04Capabilities Matrix

One brain across five ad ecosystems

Each row is a channel. Each column is a layer of the stack. Color = automation depth.

Channel
Bidding
Creative
Audience
Reporting
Data Source
Automation
Google AdsSearch · PMax · DGen · Video
Ads API · GA4 · S2S
95%
Meta AdsFacebook · Instagram
Marketing API · CAPI
90%
TikTokPerformance · Spark
Business API · Events 2.0
80%
SnapchatSnap Ads · AR
Marketing API · CAPI
70%
Bigo AdsLikee · imo · Bigo Live
Bigo API · S2S postbacks
60%
Manual Assisted Co-pilot Autonomous
05Trajectory

Twelve years, one upward curve.

2025 Present

Head of Performance Marketing

MobileArts  ·  Berlin

Leading the performance org & the AI-automation roadmap. Productizing the agentic stack across MVAS, mobile & SaaS verticals.

StrategyAI roadmapHiring
2023→2025

Performance Marketing Team Leader

MobileArts  ·  Berlin

Built the team. Built the framework. Shipped the first agentic media-buying copilot internal-first, then to operators.

TeamFramework
2020→2023

Senior Performance Marketing Specialist

MobileArts

Owned multi-million-dollar budgets across 30+ markets. Set up the data pipeline that the agents now run on.

Scale-upPipelinesGeo expansion
2019→2020

Media Buyer

MobileArts

Hands-on media buying across Google & Meta. Earned the keys to the larger budgets the hard way — day-by-day optimization.

Hands-onOptimization
2018→2019

Digital Account Manager

Social Lab

Client-side digital strategy. Translated business goals into channel plans & weekly performance narratives.

StrategyClient
2013→2017

Marketing & Sales Specialist

Phoenix 4

The years that built the muscle. Direct sales, B2B marketing, learning the customer one conversation at a time.

FoundationsB2B
06Credentials

Education, certifications & languages.

Education

  • 2019 — 2023
    Bachelor of Business Administration
    Lebanese International University
  • 2013 — 2018
    BBA, Marketing
    Tishreen University

Certifications

  • Unlock Your Team's Creativity
    Leadership · Coursera
  • Building Creative Organizations
    Leadership · Coursera
  • Designing Growth Strategies
    Strategy · Coursera
  • Setting Business Unit Goals
    Strategy · Coursera

Languages

  • Arabic Native
  • English Full Professional
  • German A2 — Learning
07Thought Leadership

Notes on the future of agentic marketing.

Long-form takes on agents, automation & the next operating model for performance marketing. Coming soon — drop your email at the bottom to get the first issue.

EssayDrafting

The dashboard is dead. Long live the agent.

Why the next generation of marketers will manage decision systems, not spreadsheets — and what that means for hiring.

~ 8 min readRead soon →
PlaybookQ3 ’26

From Key-Action to Key-Outcome.

A practical guide to defining the small set of bounded actions your agent is allowed to take — and how to evolve them safely.

~ 12 min readRead soon →
Case StudyQ4 ’26

How we ran 30 markets with a team of four.

Behind the scenes of the agentic stack at MobileArts: where humans add value, where they shouldn't, and what broke first.

~ 15 min readRead soon →
08Let's talk

Building an agentic stack for your performance team?

I take on a small number of advisory engagements & full-time leadership conversations each quarter. Tell me about the system you want to build — I'll tell you where the leverage is.

Email
fayez.abouhassoun@gmail.com
LinkedIn
/in/fayez-abou-hassoun
Based in
Berlin · CET / Remote-friendly
BERLIN · CET
--:--
—
Book a call
Q2 ’26 — 1 advisory slot
Q3 ’26 — 2 advisory slots
FT roles — selective