Ecom Automation Expert and Controller (Social Media, Crossposting, AI Video Creation, Creative Briefs)

# Job Offer: Senior AI Social Automation Architect

### Build the operating system behind a multi-market AI content network

We are building a next-generation organic social commerce engine across multiple European markets.

This is not a social media manager role.

This is not a “make a few automations in ---------- ” role.

We are looking for someone who can architect, build and continuously improve an AI-native system that turns:

**Product → Audience → Pain Point → Content Idea → Script → Video → Localisation → Publishing → Engagement → Funnel → Revenue → Learning**

into a scalable automated workflow.

Our initial focus is Finland, Sweden, Poland, Netherlands and Belgium, with more markets to follow.

The long-term goal is to operate dozens — potentially hundreds — of localised social accounts across Instagram, TikTok, YouTube Shorts and Facebook.

---

## Your Mission

Build the infrastructure that allows us to launch and operate social brands at scale.

A typical setup may look like:

**3 markets × 3 accounts × 3 posts per day**

That already means:

**27 localised videos per day**

Across multiple platforms, this quickly becomes thousands of monthly content placements.

Your job is to make that scalable, measurable and increasingly autonomous.

---

# What You Will Build

You will design the entire automation layer connecting AI models, APIs, databases, social platforms, video-generation tools and analytics systems.

A typical workflow might look like:

**Market / Product / Persona**

→ Pain Point Research
→ Content Angle
→ Hook Generation
→ Script
→ Localisation
→ Avatar / Creator
→ Voice
→ B-roll / AI Video
→ Automated Editing
→ Captions
→ QA
→ Publishing
→ Engagement
→ Analytics
→ Funnel Revenue
→ Performance Feedback Loop

The system should become smarter over time based on what actually performs.

---

# The Content Engine

We do not want generic AI scripts.

The system needs to understand:

### Pain

What keeps the target audience frustrated?

### Desire

What do they actually want their life to look like?

### Identity

Who do they want to become?

### Transformation

What change are they hoping to experience?

For example, a 60+ wellness account is not really selling:

**protein, supplements or recipes.**

It may be selling:

**energy, confidence, attractiveness, independence, longevity and the feeling of still being fully alive at 60+.**

The scripting system needs to understand that distinction.

---

# Content Formats

You will help create automated pipelines for formats such as:

* Lifestyle creators
* Wellness creators
* Fitness 50+
* Healthy ageing
* Beauty
* Recipes
* Nutrition
* Biohacking
* Educational content
* Product education
* Storytelling
* UGC-style content
* Animated / 3D characters
* AI hosts
* Real creator digital twins
* Faceless accounts

The objective is not overly polished AI content.

The objective is:

**Believable. Native. Emotional. Interesting. Consistent.**

Sometimes an imperfect-looking smartphone-style video will outperform a visually perfect AI commercial.

You need to understand that.

---

# AI Avatar Infrastructure

A major part of the role will involve creating consistent digital creators.

For example:

A female wellness creator in her 60s.

The character needs consistency across:

* Face
* Body
* Age
* Hair
* Voice
* Clothing
* Personality
* Environment
* Communication style

She should look like the same creator after 300 videos.

The same applies to animated characters and educational personalities.

Where fictional or AI-generated creators are used, the system should follow applicable platform disclosure and advertising requirements rather than relying on deceptive impersonation.

---

# Localisation

We are not interested in simply translating English scripts.

Each market should feel native.

The system should handle:

**Master Concept**

→ Finnish
→ Swedish
→ Polish
→ Dutch
→ Flemish
→ Future markets

Localisation may include:

* Native phrasing
* Local cultural references
* Localised hooks
* Local voices
* Local products
* Local pricing
* Local currencies
* Local landing pages
* Local CTAs
* Market-specific pain points
* Different creative styles

We want:

**Central strategy. Local execution.**

---

# Competitive Intelligence

We want systems that continuously analyse successful public social content in our niches.

The objective is not to copy creators.

The objective is to understand patterns.

For example:

Analyse hundreds of high-performing videos.

Extract:

* Hooks
* Topics
* Video structures
* Pain points
* Story formats
* First frames
* Video lengths
* Captions
* CTAs
* Comments
* Audience questions
* Emerging trends

Then cluster these into strategic insights.

Example:

**500 videos**

→ transcription
→ classification
→ hook analysis
→ pain-point clustering
→ engagement analysis
→ original strategy generation

Tools may include:

* Apify
* platform APIs
* social intelligence software
* transcription tools
* LLMs
* custom scrapers where permitted
* databases
* internal dashboards

---

# Automation Stack

We do not care which tool is fashionable.

We care about:

**Quality × Cost × Reliability × Speed × Scalability**

Potential technologies may include:

### Automation

* n8n
* Make
* custom workflows
* event-driven services

### AI

* Claude
* Gemini
* OpenAI models
* specialised models

### Video

* Higgsfield
* Veo
* Kling
* Runway
* HeyGen
* emerging video models

### Images

* Gemini Image
* Flux
* specialised character models

### Voice

* ElevenLabs
* other high-quality multilingual TTS

### Rendering

* Creatomate
* Remotion
* FFmpeg
* Shotstack

### Internal Tools

* Base44
* Supabase
* PostgreSQL
* custom dashboards

### Distribution

* Metricool
* Ayrshare
* ----------
* official platform APIs

### Agent Infrastructure

* MCP
* Claude Code
* custom agents
* API orchestration

You should continuously test whether something better has appeared.

---

# We Want Agents, Not Just Workflows

Eventually, we want to issue instructions such as:

**“Prepare tomorrow's three videos for the Finnish wellness account.”**

The system should then:

1. Review previous performance.
2. Find winning themes.
3. Select pain points.
4. Select content angles.
5. Generate scripts.
6. Localise them.
7. Generate assets.
8. Create voice.
9. Assemble the videos.
10. Run QA.
11. Queue the content.
12. Publish.
13. Collect performance data.
14. Update the content database.

Humans should increasingly manage exceptions and strategy instead of manually moving files between tools.

---

# Publishing Infrastructure

You will build reliable distribution across:

* TikTok
* Instagram Reels
* Facebook Reels
* YouTube Shorts

The system should account for platform-specific requirements such as:

* Video length
* Captions
* Titles
* Descriptions
* Cover frames
* Subtitle positioning
* Safe zones
* CTAs
* Posting schedules
* Aspect ratios
* Content formatting

One video does not necessarily mean one identical upload everywhere.

---

# Engagement Automation

You will also investigate intelligent engagement workflows.

For example:

**New comment**

→ understand intent
→ classify comment
→ generate appropriate response
→ automatically answer suitable FAQs
→ escalate sensitive comments
→ identify buying intent
→ identify customer-service issues

The goal is to assist genuine account engagement at scale, not create spam or fake engagement.

---

# Account Management

As we scale, we need a control centre for the entire network.

Every account should have structured information such as:

* Market
* Language
* Persona
* Product
* Funnel
* Platform
* Posting status
* Last post
* Next post
* Account owner
* Performance
* Funnel clicks
* Revenue
* Errors
* Automation status

We may operate accounts through dedicated devices or managed mobile infrastructure where appropriate.

The objective is stable, secure multi-account operations — not bypassing platform enforcement systems.

---

# Automated QA

The system needs to catch bad content before it goes live.

Examples:

* Broken subtitles
* Wrong language
* Wrong CTA
* Wrong landing page
* Broken lip sync
* Character inconsistency
* Strange AI artefacts
* Incorrect product
* Unsupported health claims
* Duplicate content
* Failed rendering
* Incorrect currency
* Watermarks
* Audio problems
* Incorrect aspect ratio

You should design automated checks wherever possible.

---

# Health & Supplement Content

Some of our products may involve health, wellness and supplements.

This requires stronger QA.

The content engine should work from approved claims and approved product information rather than allowing models to invent medical or health promises.

The automation architecture needs to support this.

---

# Performance Feedback Loop

This is one of the most important parts of the job.

Every video should become structured data.

Example:

**Market:** Poland
**Persona:** Women 55+
**Pain:** Low Energy
**Angle:** Identity
**Hook:** 07
**Format:** Lifestyle Story
**Product:** X

Performance:

* Views
* Retention
* Completion rate
* Saves
* Shares
* Comments
* Profile visits
* Link clicks
* Purchases
* Revenue

The system should eventually identify:

> “Identity-based hooks about energy significantly outperform educational hooks for this persona in Poland.”

Then create more content around that pattern.

Our competitive advantage should eventually become our own dataset.

---

# We Care About Revenue

Follower count is useful.

Views are useful.

But they are not the final objective.

We care about:

### Revenue per 1,000 views

### Funnel CTR

### Revenue per account

### Revenue per content angle

### Revenue per persona

### Revenue per market

### Customer acquisition economics

The infrastructure needs to connect social performance to commercial performance.

---

# Your Daily Responsibilities

You will monitor and improve the machine every day.

This includes:

* Check failed workflows
* Check failed uploads
* Check API issues
* Review AI output
* Review account health
* Monitor content quality
* Monitor generation costs
* Monitor localisation quality
* Review analytics
* Improve prompts
* Improve scripts
* Test new tools
* Test new models
* Reduce unnecessary costs
* Improve speed
* Reduce manual intervention
* Fix edge cases
* Improve dashboards

You are responsible for the **system**, not merely individual automations.

---

# Weekly Responsibilities

Every week we expect a short operating review.

### What worked?

Which:

* Hooks
* Markets
* Personas
* Formats
* Products
* Characters
* Angles

performed best?

### What failed?

What broke?

Why?

### What can be automated next?

Where are humans still wasting time?

### What changed in AI?

What new model or tool should we test?

### Where can costs drop?

Can we achieve the same output for 30% less?

### What do we test next week?

There should always be an experimentation roadmap.

---

# Ideal Candidate

You are probably someone who already spends too much time experimenting with new AI tools.

You should be comfortable with:

* n8n
* APIs
* REST
* Webhooks
* MCP
* OAuth
* JSON
* JavaScript
* Python
* Databases
* SQL
* LLMs
* Prompt systems
* AI video
* AI images
* AI voices
* Browser automation
* Social APIs
* Cloud services
* FFmpeg / rendering workflows
* Debugging
* Analytics

You do not have to be a hardcore backend engineer.

But you need to be extremely good at **making software work together.**

---

# We Are NOT Looking For

Someone whose entire automation skill set is:

**ChatGPT → ---------- → Google Sheet.**

Someone who recommends tools without understanding API limitations.

Someone who builds workflows that work once and break every third day.

Someone obsessed with AI quality while ignoring generation cost.

Someone who understands automation but has zero understanding of marketing.

Someone who celebrates views without asking whether they generated sales.

---

# You Think Like an Operator

When you see repetitive work:

> “Why are we doing this manually?”

When you see an expensive model:

> “Does the performance justify the cost?”

When you see a new AI tool:

> “Can I test this against what we're currently using?”

When something breaks:

> “How do I make sure this failure can never happen silently again?”

When content performs:

> “What caused it, and how can we systematically reproduce it?”

---

# First 30 Days

Your first assignment is to build the V1 operating system.

### Initial target

**3 countries**

×

**3 accounts per country**

×

**3 posts per account/day**

=

## 27 localised videos per day

The first version should automate as much as possible of:

1. Research
2. Pain-point collection
3. Content ideation
4. Hook generation
5. Script generation
6. Localisation
7. Voice generation
8. Avatar/content generation
9. B-roll
10. Editing
11. Subtitles
12. QA
13. Publishing
14. Analytics collection
15. Performance storage

---

# 60-Day Goal

By Day 60:

* Significantly less manual work
* Stable content pipeline
* Reliable publishing
* Centralised dashboard
* Better creative consistency
* Automated performance collection
* Basic winner identification
* Automated content variation
* Clear production economics

---

# 90-Day Goal

By Day 90, we want to move toward:

**Product + Market + Persona**

### IN

↓

Research
Scripts
Creative
Localisation
Publishing
Analytics

↓

### CONTENT ENGINE OUT

The team should mainly focus on:

**Products, strategy, creative direction and scaling.**

Not moving files between SaaS tools.

---

# What Success Looks Like

The best person in this role will eventually be able to say:

> “We launched three new countries this week without adding another employee.”

That is the standard.

---

# Why This Role Is Interesting

You will not be maintaining a boring internal automation.

You will be building an infrastructure where:

**AI + Content + E-commerce + Automation + Data**

all meet.

You will have significant freedom to choose the technology.

If a better model launches tomorrow, we want to know.

If a workflow can be replaced with an agent, test it.

If we can eliminate a SaaS subscription with a small internal tool, build it.

If a €500/month tool produces €20,000/month of additional output, use it.

We want someone who thinks commercially, not dogmatically.

---

# Application

Do not send us a generic CV only.

Show us what you can build.

Please send:

**1. Your strongest automation project**

Explain what it did and how the architecture worked.

**2. Your current preferred AI content stack**

What would you use today for:

* scripting
* localisation
* image generation
* video generation
* avatars
* voices
* editing
* publishing
* analytics

And why.

**3. Architecture Challenge**

Explain how you would build:

> **3 markets × 3 accounts × 3 videos/day across TikTok, Instagram and YouTube Shorts.**

We want to understand:

* Architecture
* Software
* APIs
* Database
* Estimated monthly costs
* Failure handling
* How you would keep characters consistent
* How you would collect performance data
* How you would improve the system over time

**4. Tell us what you would NOT automate.**

This question matters.

---

## Role

**Senior AI Social Automation Architect / AI Growth Systems Engineer**

**Remote**

**Long-term**

**High ownership**

Compensation depends on experience and demonstrated ability.

For the right person, this is not a small freelance automation project.

**You will own the technology layer of a system we intend to scale across markets, products and potentially hundreds of social channels.**