TM Rental · Mobility
TM Rental — €1.2M new ARR from a fleet ops co-pilot
Built an internal AI co-pilot for TM Rental's fleet team, then turned the playbook into a productised offering for new markets. New ARR landed at €1.2M in six months.
Outcome
Built an internal AI co-pilot for TM Rental's fleet team, then turned the playbook into a productised offering for new markets. New ARR landed at €1.2M in six months.
0
The challenge
TM Rental's fleet ops team was drowning in tickets. Damage reports, route disputes, customer complaints — every one needed a human read of contract, photo evidence, and chat history. Average resolution time was three days and trending up.
What we did
- Built a Claude-based co-pilot trained on three years of resolved tickets.
- Plugged it into Zendesk so every new ticket lands with a recommendation, cited evidence, and a draft response in the agent's voice.
- Iterated the eval harness weekly so quality didn't drift.
- Rolled the same system into two new markets as a productised offering, generating new ARR.
The outcome
- Resolution time down from 3 days to 6 hours.
- Agent satisfaction scores up 28 points (NPS-style).
- €1.2M new ARR in six months from the productised version.
- The co-pilot now writes ~40% of agent first-responses unedited.
"It's the first piece of AI that didn't make us spend more time training than working." — Head of Customer Experience, TM Rental