Portrait of Hedi Berrabha
Hedi BerrabhaGrowth, GTM and lifecycle · near Geneva

I build growth systems that run on AI.

Lifecycle, outbound, affiliates, ABM. I pick the platform, design the workflows, and run them with agents I build.

Get in touchLinkedIn
€1M+in contracts via ABM
2,000+affiliates recruited
200+campaigns migrated
10Mfollowers built
Built growth for
AlohiScalewayMonisnapPrismElgatoPSGSchalke 04GotagaQuersus

Work

Four companies, one thread.

The highlights of each role. Click See more for the detail.

CRM and lifecycle · AlohiLaunched the CRM and an affiliate programme from zero.
  • 200+ campaigns migrated with zero downtime
  • 20+ lifecycle workflows, live deliverability monitoring
  • 2,000+ affiliates recruited
  • Backed sales on outbound: set up the platforms, built the AI agents
Demand generation and ABM · ScalewayAccount-based campaigns that fed signed contracts.
  • 200K accounts across 10 industries
  • €1M+ in contracts from fintech accounts
  • Owned HubSpot: scoring and data hygiene
Growth and CRM · MonisnapFull-funnel growth across four markets.
  • +9% acquisition in FR, UK, ES and IT
  • +15% lifetime value, ROI up 6%
  • Push and in-app for each community
Founding team, first hire · PrismBuilt a gaming label and its marketing agency.
  • 10M followers across channels and talents
  • Campaigns with Elgato, Quersus, Gotaga, PSG, Schalke 04
  • €800k revenue in two years

AI at work

My agents absorb 60% of the workload. I keep the strategy.

I built them, I run them and I check every output. They take on the prep, checks and reports, so 40% of my time goes to strategy. Here are the biggest uses.

60%of the workload absorbed
60Claude Code skills
50+scheduled agents

AI in my life

Same habit, off the clock.

Done

My first apartment

Yields compared town by town, loan stress-tested, bought below asking.

Live

Spikely

A paper-trading game built on real market news.

spikely.cloud

In progress

My AI avatar

An AI twin that posts daily and sells affiliate tools.

Next up

Yield hunter agent

Scans listings daily, pings me only when the numbers work.

Beyond work

Four things that make me me.

Travel
I like to see a place from every side: Mexico, Colombia, China, Europe, Africa

Fun factIn Mexico I went from paradise beaches and chic spots to wilder, more adventurous places. Same in China. And I travelled across Algeria solo.

Culture
Literature, poetry, museums

Fun factMy prépa was economics and maths, with plenty of literature and philosophy on top. I loved both sides, and still do: a poem or a KPI, same pleasure.

Sport
Football, boxing, running

Fun factAt 14, my team won the regional cup against Olympique de Marseille.

Family
First, always

Fun factI grew up with three wonderful women. I learned early who really runs things.

The pyramid of Chichén Itzá under a blue sky with large clouds
Chichén Itzá, Yucatán
Sunset over the sea behind a wooden fence
Yucatán peninsula, November 2026

Contact

Let's talk.

hedi.berrabha.pro@gmail.com
LinkedIn

2025 to today · Alohi

CRM and lifecycle

200+campaigns migrated
20+lifecycle workflows
2,000+affiliates recruited
  1. Chose and launched Iterable as the CRM, with backend and data engineering: event schema, API integration.
  2. Migrated 200+ campaigns with no interruption in sends, then built 20+ lifecycle workflows.
  3. Built live deliverability monitoring: alerts on volume caps, bot waves, domain reputation, spam and bounces.
  4. Set the send rules that let other teams publish email and web content on their own.
  5. Launched the affiliate programme on PartnerStack from zero: 2,000+ affiliates recruited, 1,500+ re-engaged, commissions moved to a flat rate.
  6. Backed the sales team on outbound: set up the technical platforms, then built the AI agents that source accounts, pick contacts and draft sequences.
  7. Built Claude agents that now handle most of my recurring work. Nothing ships without my review.

2024 to 2025 · Scaleway

Demand generation and ABM

200Kaccounts targeted
10industries
€1M+in signed contracts
  1. Ran account-based campaigns across 10 industries and 200,000+ accounts.
  2. In fintech, 10% of targeted accounts became marketing-qualified leads, feeding €1M+ in signed contracts.
  3. Owned HubSpot Marketing Hub end to end: custom properties, lead scoring, data hygiene.
  4. Built lifecycle workflows for self-serve, sales-led and enterprise customers of the cloud and AI portfolio.

2023 to 2024 · Monisnap

Growth and CRM

+9%acquisition
+15%lifetime value
+6%ROI
  1. Ran full-funnel, multi-channel growth across France, the UK, Spain and Italy.
  2. Owned push and in-app on Customer.io, with segments and send times set per market.
  3. Segmented diaspora audiences by money corridor and life stage, and fed the findings to product.
  4. Produced and iterated paid and organic creatives (static, motion, UGC), tuned on hook rate, hold rate and ROAS.

2020 to 2023 · Prism

Founding team, first hire

10Mfollowers built
€800krevenue in two years
$175kclient ad budget
  1. First hire of a gaming label whose own creators were the storefront, and of its marketing agency.
  2. Agency side: ran websites, email (Klaviyo, HubSpot) and paid campaigns for Elgato and Quersus on a $175k budget.
  3. Co-branded collections with Gotaga, PSG and Schalke 04.

Email and lifecycle

Translate once, sync 13 languages

200lifecycle emails
13languages each
2,600versions kept in sync
  1. A change lands in the English master (copy, offer, link).
  2. The agent finds the same block in each of the 12 language versions.
  3. It rewrites only that block, in the tone already used in that language, leaving personalisation variables and tracked links untouched.
  4. Each version is rendered and checked against the English one: same structure, same links, nothing missing.
  5. Anything uncertain is flagged for review; the rest goes live.

Affiliate programme

Recruit affiliates on autopilot

2,000+affiliates recruited
300contacts a week
100%sends approved by me
  1. Profile: what my best affiliates have in common (site type, audience, content format) becomes the ideal partner profile.
  2. Discovery by competitor mention: sites that already review competing products are pre-qualified, found through referring-domain data and free search queries.
  3. Discovery on YouTube: channels kept only if they posted recently, with contact emails validated at mail-server level.
  4. Vetting: domain rating, organic traffic, backlink profile, plus a check of the traffic the partner claims.
  5. Dedup against existing partners and anyone who already wrote to us.
  6. Weekly round: 100 follow-ups first, then 200 invitations on the best segment, paused before every send for my approval.

Outbound

Outbound copy that grades itself

80/100minimum score to send
3rewrite rounds max
0sends without my go
  1. Source an account list, then rebuild company chains so one group is one account.
  2. Exclude existing customers and anyone on the exclusion lists.
  3. Score each account and confirm on its own website that it is a real fit.
  4. Pick the best person and confirm their role on LinkedIn.
  5. Write 3 emails, score them out of 100, research and rewrite until above 80, 3 rounds max. Otherwise it comes to me as "to review" with what is missing.
  6. Nothing is written to the sales tool or sent without my explicit go.

Reporting and planning

Reporting and planning on autopilot

Dailyinbox recap and standup
Weeklyreport to top management
0emails sent on its own
  1. Reads every email that arrived since its last pass, so nothing is read twice or skipped.
  2. Sorts them by channel and urgency into one recap, and drafts a reply for each email that needs one. Drafts only: it never sends.
  3. Builds my task report and daily standup from my notes and the tasks I closed.
  4. Drafts the weekly marketing report for top management, so leadership sees what shipped without asking.
  5. Plans my week and keeps my calendar in sync with my task list.

Real estate

My first apartment, with an AI analyst

  1. Claude compared rental yields town by town across my region.
  2. It rebuilt the numbers from real listings: rent, charges, tax.
  3. It stress-tested the loan against vacancy and rate scenarios.
  4. I bought a furnished rental where the numbers held, below the asking price.

In progress

My AI avatar, the influencer

  1. A digital twin of me, generated with AI.
  2. Posts every day, recommending tools I actually use through affiliate links.
  3. Scripts, editing and publishing run as an AI pipeline, reviewed by me.
  4. Goal: a creator business that runs without me on camera.

Next up

Yield hunter agent

  1. Every morning, it pulls new listings in the towns I target.
  2. For each one it estimates rent, charges and tax, then the net yield.
  3. It stress-tests the loan against vacancy and rate scenarios.
  4. It pings me only when a listing beats my target. Everything else stays silent.