Khaled Tolpa

Four years on the other end of the conversation: the question before the order, the confirmation, and whatever went wrong after delivery. In Arabic, in English, and now in German.

German B2 English B2 Arabic native

Auf Deutsch

Deutsch habe ich nicht im Klassenzimmer gelernt, sondern im Gespräch mit Muttersprachlern. Deshalb funktioniert es auch am Telefon, wenn jemand schnell spricht oder mitten im Satz das Thema wechselt.

Seit vier Jahren betreue ich E-Commerce-Kunden in Ägypten, den USA und Palästina. In kleinen Betrieben schreibt dieselbe Person den Text, den der Kunde liest, und beantwortet danach seine Fragen. Dazu kamen Kaltakquise-Anrufe auf Englisch bei Firmen in den USA. Dabei habe ich gelernt, eine schlechte Liste von einem schlechten Zeitpunkt und einem schlechten Angebot zu unterscheiden.

Ich suche eine Stelle im deutschsprachigen Kundenservice, in der beides zählt: die Sprache und die Erfahrung auf beiden Seiten des Gesprächs.

What I have actually handled

Customer service, ecommerce · 2022 to now

The whole journey, not one ticket type

The setup
Small operations, where the same person writes what the customer reads and then answers the customer who replies to it. No queue to hand things off to.
Where it broke
Most drop-off did not happen at the checkout page. It happened in the gaps: an unanswered question before the order, a confirmation call that never came, a delivery problem that nobody followed up.
What I did
Mapped the inbound touchpoints across the whole journey, then changed how each moment was handled: the question before the order, the confirmation, the problem after delivery, and the follow-up in between. In a cash-on-delivery market that work is not a courtesy, it decides whether an order becomes revenue.

Outbound B2B calling, United States · in English

What a failed call actually tells you

The setup
Calling accounting and bookkeeping firms in the United States, in English, from a script I wrote myself.
The problem
Most people answer a run of rejections by calling more, or by talking louder. Neither tells you anything, and both waste the list.
What I did
Treated every failed call as data with a cause. A call can die at the list, at the gatekeeper, at the offer, at the message, or at how well I had researched the person before dialling. Those need different fixes, so I changed one at a time and logged the outcome. The pitch went through four versions that way.

The volume was small. What came out of it was a way of listening to a call that transfers directly to a service queue, where the useful question is also why this person is unhappy rather than how fast I can end the conversation.

The scripts themselves

Not a description of a method, the method

Two of the call scripts I wrote and kept revising, cleaned of anything client-identifying and open in your browser. The first runs the same call structure across seven industries, so you can see what changes and what does not. The second is a discovery-led call for a single sector, where the questions get sharper because the problems are known in advance.

They are worth opening for the branching rather than the wording. What a call does when the answer is no is the part that decides whether the next one goes better.

Seven industries

One call structure, seven versions of it

Gatekeeper, opener, discovery, reframe, then the ask, with its own objection tree behind it. Switch industry at the top and watch the same skeleton change its questions. Written for the phone, not for a visit.

Open the script

One industry

Discovery-led call, bookkeeping firms

The same structure aimed at a single sector, where the questions get sharper because the problems are known in advance.

Open the script

Offer and order value

Selling more without pushing

On one product I priced the tiers so that net profit per order roughly doubled between a single unit and three, then matched the add-on to what the customer already had in the basket rather than offering whatever had the highest margin. That is the difference between an upsell that helps and one that makes someone regret the call.

Systems, and why a service person built them

A queue is only as good as the record behind it. If the customer's history is not in front of the person answering, the same problem gets explained three times. I have designed those records for whole businesses: what gets stored, who can see it, and which numbers tell a supervisor the process is slipping.

Two working prototypes

Customer records, process and the numbers on top

Both open in your browser, in Arabic. Each shows the workflow analysis first, so you can see what the design was reasoning from, then the screens it produced.

Healthcare

Private hospital

Patient records, approvals, bookings and scheduling, with the operational numbers a supervisor reads on top. A regulated operation where a missing approval number is a bill nobody can collect.

Open the prototype

Retail · Saudi Arabia

Clothing retail chain

Suppliers, items, costing, payments and clearance pricing. Built around seasonality and the cash problem that arrives with it, where last season's stock is this season's liquidity.

Open the prototype