Baku, September 24, AZERTAC
“The goal should be to use AI to make people more effective in their roles, not simply to remove people from the process,” said Adem Ibrahim, an AI strategist and Founder of the World AI Federation, in an exclusive interview with AZERTAC.
- What are the main risks and limitations involved in integrating the Generative Pre-trained Transformer (GPT) models into a company’s operational processes?
- From my perspective, the main risks are inaccurate or inconsistent outputs, data privacy, security, weak governance and becoming overly too reliant on AI. There is also a risk of automating a process before fully understanding whether AI is actually the right solution in the first place. GPT models should be introduced with clear controls, defined human accountability and oversight and a measurable business case.
-How can the best balance be struck between automating tasks and maintaining the human element in customer service?
- I believe automation works best for repetitive, predictable and low-risk tasks, such as handling common quick queries, summarizing information or supporting agents with faster access to knowledge. The human element should remain for complex complaints, escalations, emotionally sensitive situations, exceptions and high-value customer interactions. The goal should be to use AI to make people more effective in their roles, not simply to remove people from the process.
- What key skills are becoming critical for IT teams in this era of rapid AI development?
- AI literacy is becoming essential, alongside automation, API integration, cybersecurity, data governance and the ability to design reliable AI workflows. IT teams also need stronger commercial and operational awareness. The most valuable skill is increasingly the ability to understand a business problem, identify where AI can genuinely add value, and then implement it in a secure, governed and successful way.