As part of the Alphawave Tech Talk series, Charl van Heerden, CEO of Saigen, shared a forward-looking vision on how AI—particularly large language models (LLMs) and automatic speech recognition (ASR)—is reshaping the call centre industry. His presentation (rooted in real-world applications through Callbi AI Labs, the recent collaboration between Callbi and Saigen) explored cutting-edge solutions that address long-standing operational inefficiencies and elevate customer experience.
Why Callbi AI Labs?
Callbi AI Labs, a joint initiative between Saigen and Callbi, was established to apply LLMs and speech technology in practical, high-value use cases for the South African call centre market. With access to over 100 Callbi customers, the lab focuses on problem-first innovation.
Watch the summary here:
Key AI-Powered Innovations
1. Automatic Call Summarisation
Manual review of lengthy call recordings is inefficient and costly. AI-powered summarisation condenses conversations into structured, digestible insights-cutting down post-call work and improving continuity across customer interactions.
Impact:
- Efficiency in handling follow-ups.
- Customer satisfaction through contextual continuity.
- Time savings of 30 seconds to 2 minutes per call.
2. Multilingual Speech Recognition
Saigen’s ASR system can automatically switch between South African languages, including English, Afrikaans, isiZulu, and Sesotho—vital for accurate transcription and analysis in South African call centre environments.
3. CRM Data Validation
A major challenge faced by call centres is the mismatch between what a customer says and what an agent records. LLMs are now being used to:
- Flag discrepancies (e.g., incorrectly captured ID or bank account numbers).
- Detect error-prone behaviour.
- Significantly reduce incorrect data entries.
4. Agent Performance Analysis
By comparing high-performing and low-performing agents’ conversations, AI can pinpoint behaviours that drive success. For example:
- LLMs identified trends even seasoned QA analysts might miss.
- They are also able to identify agent behaviours which are more likely to lead to successful outcomes in the debt collection industry.
5. Scorecard Automation
Manual call evaluation is time-consuming. AI now helps to:
- Auto-fill scorecards based on transcripts.
- Standardise evaluations without human bias.
- Scale quality assurance across thousands of calls.
6. Agent Assist: Real-Time Support
Perhaps the most exciting development is the Agent Assist prototype, a live assistant that:
- Monitors calls in real time.
- Tracks compliance items as they're spoken.
- Prompts agents with helpful responses or policy info.
- Alerts managers to high-risk calls in progress.
This system is designed to enhance—not replace—human agents, particularly new staff navigating complex scenarios like prescribed debt.
“We’re beyond automating workflows, we’re amplifying people,” says Dr Barnard. “Agent Assist is about supporting humans with AI, not replacing them. This is the future of customer service: smarter, faster, and more emotionally aware.” - Dr Etienne Barnard, TechFinancials
The Future of AI in Call Centres
Charl cautioned against the overhype surrounding general AI but emphasised the practical value of today’s tools. While fully automated virtual agents may become more common, especially for repetitive inbound queries, there remains a critical role for human empathy and complex decision-making.
The Callbi AI Labs approach combines deep local language expertise, ethical data use, and tight client feedback loops—a recipe not just for innovation but for staying competitive against global giants.
Closing thoughts
Charl’s talk was both a technical showcase and a strategic outlook. The Callbi-Saigen collaboration demonstrates how AI can drive measurable impact in customer service, compliance, training, and operations—without losing the human touch.
If you’re in the call centre space and exploring how to responsibly and effectively embed AI, this is the benchmark to watch.
For more information on Callbi AI Labs, read the news feature here or contact us at info@saigen.co.za.
