The idea was quickly turned into a test-ready product without a long and expensive conventional development cycle.
YL
01 / AI advertising video platformYevhen L.
Founder
AI • AUTOMATION • CONTENT
Automation that works as a system.
I design AI integrations, content pipelines and personalized outreach systems with n8n, Make and OpenAI.
7detailed case studies10+connected workflowsHITLquality control
ORCHESTRATIONn8n / AI / Logic
AIAI analysisOpenAI / GPT
TXTContent generationPosts / Email
@Email outreachPersonalization
INInbound leadsForms / Webhook
DBCRM / databaseLeads / Companies
ICPSegmentationAudiences / ICP
↗AnalyticsReports / KPI
DATAStorageFiles / Data
Apify
n8n
OpenAI
Airtable
SELECTED WORK / 2026
Projects where AI becomes part of the process
Not a standalone button or chatbot, but a working system with data, checks and measurable outcomes.
TIMELINE2.5 weeksOUTCOMEMVP readyConcept→UX→AI tools→MVP✓MVP for an AI advertising video platform
A production-ready MVP for a US AI startup with multi-role access, payments and asynchronous advertising video generation.
Concept → architecture → AI generation → interface → working MVP
TIMELINE2.5 weeksOUTCOMEMVP ready
Concept→
UX→
AI tools→
MVP
✓MVP for an AI advertising video platform
A production-ready MVP for a US AI startup with multi-role access, payments and asynchronous advertising video generation.
Concept → architecture → AI generation → interface → working MVP
01 / CHALLENGE
Validate an AI marketplace concept quickly, without a long and expensive conventional development cycle.
02 / SOLUTION
Lovable powers the interface, n8n orchestrates the backend, and Supabase handles data, authentication and storage. Stripe, Veo, NanoBanana and KIE are integrated.
03 / OUTCOME
A fully functional MVP with roles, payments, asynchronous media generation and scalable SaaS architecture.
View on Upwork ↗
SYSTEMS2 CRMsTIME SAVEDup to 40 h/moSalesforce→Mapping→Sync→Agile CRM✓Salesforce → Agile CRM synchronization
A complete integration synchronizes companies, contacts and deals between two CRMs, removes manual entry and creates one reliable data flow.
Salesforce → processing and mapping → Agile CRM → monitoring
SYSTEMS2 CRMsTIME SAVEDup to 40 h/mo
Salesforce→
Mapping→
Sync→
Agile CRM
✓Salesforce → Agile CRM synchronization
A complete integration synchronizes companies, contacts and deals between two CRMs, removes manual entry and creates one reliable data flow.
Salesforce → processing and mapping → Agile CRM → monitoring
01 / CHALLENGE
Eliminate manual transfers, unsynchronized contacts and companies, duplicate records and errors between sales and marketing systems.
02 / SOLUTION
Implemented API integration, optional two-way logic, field mapping, duplicate detection, deal-stage synchronization, logging and error monitoring.
03 / OUTCOME
Data now synchronizes near real time, manual entry is nearly eliminated and the system can save up to 40 hours per employee each month.
View on Upwork ↗
TIMELINE4 weeksLEADS150+ / weekMaps→Website→AI review→HubSpot✓Partner CRM for a manufacturing company
A B2B system finds prospective partners, analyzes their websites and pre-qualifies companies before sending them to HubSpot.
Google Maps → websites → AI analysis → qualification → HubSpot
TIMELINE4 weeksLEADS150+ / week
Maps→
Website→
AI review→
HubSpot
✓Partner CRM for a manufacturing company
A B2B system finds prospective partners, analyzes their websites and pre-qualifies companies before sending them to HubSpot.
Google Maps → websites → AI analysis → qualification → HubSpot
01 / CHALLENGE
Scale B2B partner discovery across Central and Eastern Europe without reviewing every company manually.
02 / SOLUTION
The system collects companies from Google Maps, analyzes their websites with AI and sends qualified records to HubSpot.
03 / OUTCOME
A flow of more than 150 pre-qualified partner leads per week.
MAKE14 scenariosSCRIPTS20+Collect→Analyze→Email→HubSpot✓B2B lead generation, email marketing and CRM
The system collects investor data, analyzes their companies and industries, selects relevant prospects and moves them into email campaigns and HubSpot.
Investors → collection and analysis → email campaign → HubSpot
MAKE14 scenariosSCRIPTS20+
Collect→
Analyze→
Email→
HubSpot
✓B2B lead generation, email marketing and CRM
The system collects investor data, analyzes their companies and industries, selects relevant prospects and moves them into email campaigns and HubSpot.
Investors → collection and analysis → email campaign → HubSpot
01 / CHALLENGE
Help a startup collect information about prospective investors, assess relevance and build a repeatable communication process.
02 / SOLUTION
Crunchbase, Apify and Google Maps supply the data; 14 Make scenarios and 20+ Apps Scripts process it, Snov.io runs campaigns and HubSpot stores the outcome.
03 / OUTCOME
The client received one automated system for lead collection, email marketing and ongoing CRM work.
View on Upwork ↗
SYSTEM10+ workflowsCONTROLHuman-in-loopImport→Enrich→AI signals→Review✓AI lead enrichment and personalization system
An end-to-end pipeline imports leads, collects signals from LinkedIn, websites, news and ads, generates personalization and checks quality before outreach.
Leads → enrichment → signals → AI personalization → review
SYSTEM10+ workflowsCONTROLHuman-in-loop
Import→
Enrich→
AI signals→
Review
✓AI lead enrichment and personalization system
An end-to-end pipeline imports leads, collects signals from LinkedIn, websites, news and ads, generates personalization and checks quality before outreach.
Leads → enrichment → signals → AI personalization → review
01 / CHALLENGE
Combine data from many sources and generate personalization without sending weak or risky variants.
02 / SOLUTION
An orchestrator manages separate enrichment workflows, signal analysis, text generation, statuses and quality checks.
03 / OUTCOME
One scalable pipeline with logs and quality control that can safely process leads in batches.
CONTENTPhoto + videoFORMATPipelineMedia→Script→Voice→Publish✓Social media content factory
Photos and videos from supplier websites and Google Drive are prepared, enhanced with AI, edited, voiced and assembled into publication-ready assets.
Media → script → editing → voiceover → publishing
CONTENTPhoto + videoFORMATPipeline
Media→
Script→
Voice→
Publish
✓Social media content factory
Photos and videos from supplier websites and Google Drive are prepared, enhanced with AI, edited, voiced and assembled into publication-ready assets.
Media → script → editing → voiceover → publishing
01 / CHALLENGE
Turn fragmented supplier assets and original footage into a consistent stream of social content.
02 / SOLUTION
Airtable stores assets and statuses, AI assists with scripts, Vizard processes video, ElevenLabs creates voiceovers and FFmpeg builds final versions.
03 / OUTCOME
A repeatable content pipeline with one database, transparent stages and assets prepared for automated publishing.
SOURCESCompany sitesDATABASEAirtableSupplier→Apify→Normalize→Airtable✓Heavy equipment catalog parsers
A set of workflows collects specifications, prices, descriptions, photos and videos from supplier websites, normalizes the data and updates the catalog.
URL → page parsing → normalization → deduplication → Airtable
SOURCESCompany sitesDATABASEAirtable
Supplier→
Apify→
Normalize→
Airtable
✓Heavy equipment catalog parsers
A set of workflows collects specifications, prices, descriptions, photos and videos from supplier websites, normalizes the data and updates the catalog.
URL → page parsing → normalization → deduplication → Airtable
01 / CHALLENGE
Automatically collect inconsistent equipment listings from multiple websites, including images and embedded videos.
02 / SOLUTION
Each supplier has a dedicated extraction workflow; shared logic then maps fields into one structure and removes duplicates.
03 / OUTCOME
An up-to-date equipment catalog without manual copying, with consistent data and source tracking for every record.
CLIENT FEEDBACK
What changed after the systems went live
These testimonials relate to the client projects above and show the impact from the customer team's perspective.
Automated data transfer reduced manual work and allowed the team to focus on the most valuable leads.
OM
02 / M&A boutiqueOctavian M.
VP of Business Development
The system created a focused process for finding and qualifying partners for the CEE growth program.
TF
03 / Manufacturing companyTony F.
Head of Partnerships
Automated outreach helped us find suitable investors and approach them personally at the right moment.
JS
04 / Pre-seed startupJames S.
Co-Founder
WHAT I BUILD
From one integration to a complete pipeline
Automation architecture
System design, connections, queues, statuses and safe error-handling workflows.
AI and personalization
Signal extraction, content generation, quality scoring and human-in-the-loop review.
Integrations and data
CRMs, spreadsheets, APIs, webhooks, parsers, synchronization and data normalization.
Content pipelines
Automated photo and video processing, voiceover, assembly and publishing.
APPROACH
Working logic first. Scale second.
Every project starts with a simple, testable workflow. Automation expands only after the core process works reliably on real data.
Define the problem
I clarify the goal, data sources, constraints and the points where a person should make the decision.
Build the prototype
I launch a minimum working workflow on real data and test the edge cases.
Turn it into a system
I add logs, statuses, error handling, documentation and a safe path to scale.
NEXT SYSTEM
Is there a process ready to be automated?
Describe the task in a few sentences and I will propose the structure of the first working prototype.