case study · 2024–2025

Just Eat UK: rider recruitment, automated end to end.

Just Eat's UK rider-recruitment team needed the whole process automated, from a rider's first message to onboarding. I designed the conversation, built the workflow in n8n and wrote the APIs that connect it to their systems.

to launch
7 months
rider flows
4
Client
Just Eat (UK), via Ciklum
Role
AI workflow integration
Timeline
Feb 2024 – Jan 2025
Stack
n8n · TypeScript · Node.js · REST APIs · Twilio · AWS · Laravel Forge · GitLab · MySQL · MariaDB

the brief

Every applicant, one workflow.

Riders apply on a bike, a cycle, an e-scooter or in a car, and each type needs different questions and different checks. Every applicant also needs a medical check and a string of follow-up messages.

The goal was simple to say and hard to build: automate the whole recruitment process, with no one needed in the loop, while handling medical data to EU public-health rules.

what I built

Watch it run.

a replay of the real flow, simplified

A rider's application triggers the workflow. An AI chat screens them, a switch sends them down the right flow for their vehicle, and the medical check gates everything after it.

Once they pass, Twilio sends the SMS and email and the rider is onboarded in Just Eat's CMS, without anyone touching it.

  1. Webhook: rider applies
  2. AI chat: screening
  3. Switch: vehicle type
  4. Medical check
  5. Twilio: SMS + email
  6. CMS: onboard rider
last run: succeeded

inside the build

Six problems, solved once.

  • A conversation, designed

    I designed the chatbot's conversation and wrote the APIs that connect it to Just Eat's CMS.

  • Four rider flows

    Separate scenarios for bike, cycle, e-scooter and car riders, each with its own questions and checks.

  • Medical data, handled carefully

    Medical reporting and data security were the main focus, built to meet EU public-health policy.

  • Messages that send themselves

    SMS and email go out through Twilio the moment a rider's medical is passed.

  • Answers to new questions

    AI-driven responses, with the model trained to answer queries the script didn't cover.

  • Routes, optimised

    Best-route finding and optimisation for riders, using Google Maps.

the result

Launched in seven months, with the whole recruitment process automated.

from brief to launch
7 mo
rider flows live
4
public-health rules designed in
EU

more work

09trigger · your_project

Thirty minutes, no slides. Tell me what's slowing your team down and I'll tell you honestly if I'm the right fit.