Customer Experience · CX Analysis · AI Automation

Push
further.

Most businesses don't have a technology problem. They have a customer journey problem and often don't know where it starts.
Understanding first. Building second.

100%
Follow-up Rate
−5h
Saved per Week
Retention Uplift

Philosophy

"You've got to start with the customer experience and work backward to the technology."

Steve Jobs, WWDC 1997

We don't chase tools. We eliminate friction in your customer journey. Only once we understand where your customers lose momentum do we deploy AI — precisely and purposefully.

About

15 years of
real customers.

I'm Gabor — and I didn't learn customer experience from a course. I lived it. For over 15 years I led teams at Nike, TK Maxx and Mister Spex — managing stores, coaching people, and obsessing over why customers come back or don't.

What I kept seeing: a customer walks in with a question — and leaves with a bad feeling because nobody had time. An email that never got answered. A process built for the business, not for the person going through it. Nobody decides to lose a customer. It just happens — in the small moments nobody is watching.

Customer satisfaction isn't a metric — it's a feeling. The businesses that get it right are rarely the ones with the most tools. They're the ones who understand where their customers lose momentum — and remove that friction deliberately. Sometimes that's a small automation. Sometimes it's just clarity about the journey itself, and how the team shows up at each step. Either way, the goal is the same: free people up to focus on what actually matters.

Before I touch a single tool, I map the customer journey — where people hesitate, where they leave without saying why. The answer isn't always a system. Sometimes it's a single conversation that changes everything.

Customer Journey CX Analysis Process Design Make.com Notion HubSpot Anthropic API Tally

"The gap between retail and digital CSM is smaller than people think. In both worlds, the goal is the same: make the customer feel understood — and give them a reason to come back."

GK

Gabor Komjathy

ATPUSH

15+
Years in Retail
Languages
100%
Customer First
0
Wasted Tool Chasing

Real Projects

Real problems.
Real solutions.

01 / Full CSM System Build · Sport Coaching

Art of Gerundium

artofgerundium.com ↗

The Situation

David, a self-employed sport coach in Hungary, had no website, no CRM, and no system. Leads came only through personal referrals. Customer communication was scattered across 5+ tools. Admin work alone consumed 5+ hours every week.

What I Built

A complete system from scratch: professional website on Netlify, lead capture via Tally embedded across website and social media bios (IG, FB, TikTok), automated into a Google Sheets CRM and Notion Hub via Make.com. Every new enquiry triggers an instant confirmation email and a push notification to David — within 2 seconds.

Next Phase

Re-engagement automation for customers inactive 60+ days, follow-up sequences after proposals, and DACH market expansion for online coaching — targeting Germany and Austria.

Make.com Tally Notion Google Sheets Netlify
100%
Response Rate
Was: manual & irregular
−80%
Admin Time
From 5h+ to <1h / week
+200%
Lead Target
From <10 to 20–40 / month
~$10
Monthly Cost
Full stack, all automated
02 / CRM & Automation System

Customer Health
Scoring System.

The Challenge

Customer risks were invisible until it was too late. There was no way to identify which customers were about to churn — until they already had. Reactive support instead of proactive care.

The Analysis

The real problem wasn't the churn — it was the silence before it. Customers were disengaging weeks before they left, but nobody was watching the right signals. The journey had a gap: between the last positive interaction and the moment it was too late to act. The system needed to close that gap — not replace human judgement, but make sure the right person was alerted at the right moment.

The Solution

Designed and built an automated Customer Health Scoring System to identify at-risk customers before they leave. Google Sheets feeds structured customer data into HubSpot via Zapier. When a health score drops below threshold, a ticket is automatically created and the team is alerted for proactive outreach.

The Result

From reactive firefighting to proactive retention. At-risk customers are identified and contacted before they churn — shifting the entire support model from response to prevention.

HubSpot Zapier Google Sheets Python
Proactive
Not Reactive
Customers flagged before churn
Auto
Ticket Creation
Zero manual monitoring
03 / AI Customer Agent · Sport Coaching

Configurable AI Agent

First deployment — Art of Gerundium

The Situation

Before any tool is chosen, there are two customers to understand: the business owner, and their clients. For a coaching business spanning three languages and two countries, both had the same need — clarity at the moment of first contact. The question was never "which technology". It was: what breaks down between a potential client reaching out and actually showing up?

What I Built

A lightweight, configurable AI agent built on the Anthropic API. It understands context, responds in the customer's language, resolves what it can — and knows when to escalate to a human. One configuration file per client. No reengineering needed.

The Approach

Analyse first. Build second. The architecture follows the business — not the other way around.

Python Anthropic API FastAPI Railway

Agent Logic

💬
Customer
Asks
🔍
FAQ
Search
🤖
Resolve or
Escalate
Right
Answer

"Trilingual support (DE/EN/HU) — resolves what it can, escalates what it can't."

Live Demo — Ask the Agent

👋 Hallo! Ich bin der AOG-Agent. Frag mich etwas über Art of Gerundium — z.B. Preise, Ausrüstung oder Kontakt.

My Process

Three steps.
One north star.

  1. 01

    Diagnose

    I map your customer journey end-to-end. Where do you lose people? What frustrates them? Which touchpoints are broken?

  2. 02

    Design

    We design the ideal experience — before touching a single tool. Humans first. Technology follows.

  3. 03

    Deploy

    I build and hand over the system — fully documented, automated, and scalable. You stay in control. Always.