Two years ago, over a coffee conversation, my colleague and I faced a stark reality: our service management division costs weren’t sustainable. We were delivering quality managed services to over 2 500 customers using ITIL 4 methodology and the SIAM service operating model. Our governance was solid, our processes mature, and our customer satisfaction was improving. But the financial math didn’t work. Something had to change.

The traditional response would have been to cut services, reduce headcount, or compromise on quality. Instead, we asked a different question: What if we could maintain or even improve service excellence while dramatically reducing operational costs?

Grow, don’t buy

Here’s where we diverged from conventional wisdom. Rather than hiring expensive AI specialists or purchasing off-the-shelf solutions, we made a bold decision: we would reskill our existing service desk agents to become our AI and automation developers.

This wasn’t just about cost savings – it was about sustainability. Who better to understand the pain points of service delivery than the people living them daily? Who better to design solutions than those who had fielded thousands of customer calls?

Today, all 27 members of our “innovation revolution” team are former service desk agents. Not a single external hire. They’ve built our entire generative AI framework from the ground up.

The results speak for themselves:

• • 90% of email-based service requests are now handled by generative AI
  • We achieved our highest Net Promoter Score in a decade
  • SLA achievement reached 98.74% across our entire customer base – a BCX record
  • One major client went from consistent SLA failures to zero failures in a year

But perhaps our most impactful innovation addresses a uniquely South African challenge: load-shedding. When power outages occur, our client would receive 5 000 calls in 30 minutes, overwhelming traditional support models. Our High Impact Incident Management system now automatically identifies affected substations and dispatches engineers within 5-10 minutes, completely automated. This isn’t just better customer service – it’s faster power restoration for thousands of South Africans.

Local is lekker

One critical insight: global AI solutions often miss local nuances. Our voice bot speaks isiZulu, isiXhosa, Afrikaans, English, and Portuguese, because South African customers deserve to be served in their preferred language. This wasn’t an afterthought; it was a core design principle.

South African customers have different interaction patterns, expectations, and communication preferences than those in other markets. Building for our specific context wasn’t just better business – it was the only way to achieve true adoption.

Perhaps most importantly, this journey proved that digital transformation doesn’t require abandoning your workforce. It requires investing in them. One graduate who joined us two years ago just won our CEO award. This year, four young women in our internship programme are building the automation framework for a major retail client’s national network.

As we look ahead, we’re not just maintaining these gains – we’re scaling them. We’re crafting a service excellence framework that balances customer needs, business sustainability, and responsible growth. The goal isn’t only operational efficiency, but also creating pathways for the next generation of technical talent.

The lesson? Sometimes the most innovative approach isn’t buying the latest technology – it’s believing in your people’s capacity to master it.