From 39 to 10 days: How Telecom-provider M-net transforms its customer service with RPA
fibre-optic provider · 850 employees · 500,000 customer connections · 150 customer service staff
| „”Using RPA is a game-changer for us: our customers benefit from faster order processing at a high level of quality, our staff can focus on more demanding tasks, and our customer service achieves excellent KPIs.” – Ferhat Datli, Head of Business Consulting, Customer Service, M-net |
The Challenge
The company’s customer service department – around 150 staff across several teams – handled new connections, tariff changes, cancellations and fault reports across four separate IT systems: document management, ERP, technical switching system and CRM. Every order required manual input in each of these systems.
The result:
| Processing times of 39 days in order management – customers waited more than five weeks for their order confirmation. |
| A significant backlog of orders, particularly in the residential segment. The backlog grew faster than it could be cleared. |
| Staff tied up by repetitive click-work across four IT systems – error-prone, frustrating, and value-destroying. |
| Rising customer complaints: “What is the status of my order?” calls clogged up the call centre. |
| Backend automation via the IT department: too complex, too expensive, too slow for operational needs. |
In short: the right idea – automation in customer service – but no framework to implement and scale it systematically.
The Approach: Four Phases to Scale
The company began with its first proof of concept using UiPath. The results were promising, but internal capacity was insufficient to enable scaling. Lunatec was engaged as a partner with a clear mandate: to build an RPA Competence Centre, scale automation beyond customer service, and ensure a lasting knowledge transfer.
| Phase 1: First POCs New connections and cancellations as the first candidates. Focus on a single network technology to reduce complexity. Phase 2: Starting to Scale Lunatec engaged. Dedicated RPA team established. Joint development: external and internal developers working side by side. Expansion to provider switches and tariff changes. Phase 3: Competence Centre 50+ processes automated, 106,000 successful transactions. Internal developers increasingly taking ownership. RPA offered as a service for the whole organisation – Finance, Planning, Operations. Phase 4: Autonomous Operations 75+ processes, 190,000 transactions. 5 unattended robots running 24/7. Internal team takes over development and operations. Lunatec as second-level support. |
Key: Knowledge transfer was not a by-product – it was a strategic objective. Career changers from customer service were trained as RPA developers through a structured mentoring programme alongside the external Lunatec developer.
The Results
The impact was measurable – and where customers feel it most: in speed. New connections that previously took 39 days are now completed in 10. Cancellations in 5 days instead of 15. The 90% end-to-end finalisation rate for new connections means 9 out of 10 orders run fully automatically, including customer communication.
The indirect effects were equally significant: fewer status enquiries in the call centre (because orders are processed faster and robots do not make mistakes), fewer cancellations (because customers no longer wait weeks), earlier revenue recognition (because orders are activated more quickly), and fewer failed engineer visits (because automated reminder SMS messages reduce the no-show rate).
What You Can Take Away
These five principles are transferable – regardless of industry, company size or level of automation maturity.
1. Start with the customer experience, not with the technology.
Processing time – from order receipt to order confirmation – was the critical metric. Not the number of automated processes, not the technology architecture. Ask yourself: which metric does your customer feel directly?
2. Scale with a Competence Centre, not with individual projects.
The company did not implement an “RPA project”. It built a Competence Centre – with its own team, its own ticket category, its own service offering for all departments. That is the difference between a pilot and a platform.
3. Train internal developers with a structured path.
Career changers from customer service were developed into fully qualified RPA developers across four phases. The key: mentoring by the external partner rather than classroom training. After Phase 4, the internal team manages development and operations independently.
4. Invest in change management before the first bot.
The works agreement was not a bureaucratic formality. It was the moment the workforce understood: no redundancies, clear rules, transparency. Without that step, acceptance would never have come – and without acceptance, no scaling.
5. Plan for operations from day one.
Five unattended robots running 24/7 require at least three full-time FTEs for monitoring, bug-fixing and further development. Anyone who does not plan for this will have unstable bots and frustrated stakeholders after six months. The M-net experience shows: a team of three FTEs is the minimum for stable operations.
NEXT STEP
Which of your customer service processes has the greatest automation potential?
In 30 minutes we will look together at your three most labour-intensive processes and show you where the greatest levers for processing time, quality and customer satisfaction lie. No sales pitch. Just an honest assessment.
| What you will take away from the discovery call: ✓ An initial assessment of your automation potential ✓ Concrete examples from your industry ✓ A realistic timeline and investment framework |