I automate processes so they run themselves. You count the results.
System > Tool
I build KPI-oriented processes, not a collection of disconnected plugins and toys.
E-commerce & Customer Service
I specialize in reducing TTR, lowering support costs, and boosting NPS in e-commerce.
Fast MVP
I deliver first measurable results in 14-30 days, then iteratively develop the system.
Mini Case Studies (Results in 30 days)
Problem: Long TTR in complaints (48h), support costs rising.
System: LLM + email auto-classification + SLA alerts + intelligent response templates.
Result (30 days): TTR ↓72%, cost/ticket ↓41%, NPS ↑+1.2.
Next: Integration with manufacturer system + auto-labeling returns.
Problem: Low lead quality and slow sales response to new inquiries.
System: AI scoring, lead data enrichment, and automatic CRM assignment.
Result (30 days): Lead quality (SQL) up 35%, response time cut from 2h to 5 minutes.
Next: Automatic follow-up email personalization based on AI scoring.
Problem: Preparing legal documents takes 4-6h. Lots of time searching for templates and clauses.
System: AI for document drafting based on facts + RAG for clause search in internal knowledge base.
Result (30 days): Document prep time ↓70% (to 1.5h), substantive errors ↓90%, template search from 30 min to 30 sec.
Next: Integration with e-court for automatic filing and case status monitoring.
I've built systems that...
Reduced average complaint handling time (TTR).
Cut the cost of handling a single ticket.
Accelerated new client onboarding.
Saved in analytical work for reporting.
Start with one problem
When is it worth it, and when is it not?
✅Good collaboration if...
- You have repetitive, rule-based processes.
- You want to make decisions based on hard data.
- You understand that technology is an investment, not a cost.
- You have a dedicated process owner on your side.
- You're ready for an iterative approach after MVP deployment.
❌It won't work if...
- You lack a dedicated process owner on your side.
- We don't have a clearly defined, measurable goal (KPI).
- The main motive is "AI because it's trendy", not a real problem.
- You're looking for a "set it and forget it" project, without iterations.
English content coming soon
We are working on English versions of our articles.