Case Studies - What I've Built
Problem:
A manufacturing company (blinds, shutters) managed orders in Google Sheets and on paper. Sales reps wasted time manually creating quotes, often made calculation errors, and the owner had no visibility into the sales pipeline.
Hypothesis:
"A dedicated CRM with product configurator and automatic pricing will reduce order handling time by 70% and eliminate calculation errors."
System:
Built a CRM from scratch in React + Node.js with modules: product configurator (blinds, shutters, mosquito nets), automatic pricing from price list, PDF quote generation, customer and order management, installer panel, BI dashboard with sales analysis. System runs on client's own server.
Result (30 days):
Quote creation time: from 30 min to 3 min. Zero calculation errors. Owner has full real-time sales visibility. ROI in 45 days.
What's next:
Integration with warehouse system, installation planning module with calendar and SMS notifications for customers.
Problem:
Shop owners and service providers spent dozens of hours writing complaint responses. They often didn't know their rights, and each complaint required research and legal consultation.
Hypothesis:
"An AI tool that analyzes complaint content and generates professional, legally compliant responses will reduce handling time by 90% and improve response quality."
System:
Created a web app with AI (GPT-4 + RAG based on consumer regulations). User pastes complaint text, system analyzes legal basis, assesses claim validity, and generates ready response with law citations. Plus: case history, templates, PDF export.
Result (30 days):
Complaint response time: from 2h to 5 min. 95% of users rate responses as 'professional'. No need for legal consultations in standard cases.
What's next:
Expand with mediation module, integration with e-commerce platforms (Allegro, BaseLinker) and automatic order data retrieval.
Problem:
Marketing agency spent ~8 hours weekly manually preparing client reports. Data was scattered across Google Ads, Facebook Ads, and Google Analytics.
Hypothesis:
"Automating data aggregation and report generation will reduce work time by 90% and eliminate errors."
System:
Created n8n workflow that daily fetches data from ad platform APIs, aggregates it, then sends to Looker Studio template. Clients get link to always-current dashboard, managers get Slack summary.
Result (30 days):
30 hours saved monthly. Reports available in real-time. 100% client satisfaction with new reporting format.
What's next:
Add AI module that analyzes data and adds automatic insights and recommendations to reports.
Problem:
SaaS company had low trial conversion. Sales team didn't know which users to focus on.
Hypothesis:
"Introducing automatic lead scoring based on in-app activity will identify most engaged users and increase conversion."
System:
Integrated app data (Mixpanel events) with CRM (Pipedrive) via n8n. Workflow assigns points for key actions (e.g., project creation, team member invite). Users exceeding score threshold are marked 'Hot Lead' and move to top of sales rep's task list.
Result (30 days):
22% increase in trial-to-paid conversion in first month. Sales team focuses on highest potential leads.
What's next:
Expand system with personalized activation emails, sent automatically based on user actions (or inactions).
Problem:
B2B consultant wanted to build personal brand on LinkedIn, but writing and publishing posts took 5-6h weekly. Lacked consistency, and posts weren't optimized for algorithm.
Hypothesis:
"AI system that generates posts from notes and auto-publishes will maintain consistency with minimal effort."
System:
Built n8n workflow integrating: Notion (ideas and notes database), AI (GPT-4 for content generation in author's style), Buffer/LinkedIn API for scheduling. System daily analyzes content calendar, generates post from note, adds appropriate hashtags, and publishes at optimal time.
Result (30 days):
Content time: from 6h to 30 min weekly. Publishing consistency: 100% (5 posts/week). 340% reach increase in 3 months thanks to consistency.
What's next:
Expand with auto-repurposing (one topic to LinkedIn post + X thread + newsletter), A/B headline testing and engagement analytics.