Aydem Energy manages multiple-fold seasonal call surges with an Azure OpenAI-powered digital assistant

Aydem Energy brings electricity to around 6 million customers across five cities in Türkiye, managing generation, distribution, and retail to keep energy flowing smoothly. Committed to delivering uninterrupted, high-quality service, it places great importance on customer satisfaction. "Our call center handles around 21,000 calls daily. To deliver excellent customer support, we have to rely on the highest quality digital solutions," shares Süleyman Tosun, IT Strategy and Governance Director at Aydem Energy.

Seasonal challenges, like hot summers or high number of travelers during vacation periods, lead to increased demand while windy winters cause disruptions. This can cause more outages and significantly higher call volumes to the contact center. In fact, 70% of inquiries are related to power outages. “During seasonal peaks, the number of customer calls increases significantly, making it hard for us to quickly scale our staff,” explains Tosun. “The pressure to respond quickly under challenging conditions also causes immense stress to our employees.”

Introducing a scalable, AI-powered solution

Recognizing that WhatsApp is widely used in Türkiye, Tosun set out to implement a generative AI assistant to deliver information at scale through the platform. Tosun and his team decided to start with a proof-of-concept (POC). "We looked at four POCs using different technologies," recalls Tosun. After reviewing the outcomes, they concluded that Azure OpenAI Service was the best choice, and Softtech was the ideal partner for implementation. "Working with Softtech felt like operating as a single, unified team. Their deep understanding of our business requirements and their proactive approach enabled us to address challenges efficiently," he explains.

Understanding that generative AI is always evolving, Tosun established a dedicated test team within the call center. ”Generative AI projects are unique because new models are constantly being released. Each update can bring smarter, cheaper solutions, so we had to test everything thoroughly,“ explains Tosun. The project was a collective effort involving key team members, including Aydem Energy’s Chief Technology Officer, Engin Kavas, Digital Energy Solutions and Data Analytics Group Manager, Osman Marangoz, Call Center Manager, Seda Küçükoğlu, and IT Manager, Özgür Özer. "Many people were involved during the project and their contributions were critical to success," acknowledges Tosun.

Boosting call center efficiency

Aydem was able to launch the AI assistant in just three months. During an outage, Aydem directs customers to WhatsApp, reducing the number of calls to the call center. The digital assistant starts the interaction by securing data protection and other regulatory approvals. It then informs the customer that they are interacting with a digital assistant that’s using generative AI and explains the tasks it can perform.

The AI assistant integrates with Aydem’s internal systems to provide customers with real-time updates. For example, it matches phone numbers with internal customer relationship management (CRM) data, providing location-specific updates. It uses WhatsApp’s location sharing feature to pinpoint malfunctions, guides customers in submitting meter readings, checks outstanding bills, and processes compensation claims. If the customer reports an urgent issue or expresses anger, the digital assistant seamlessly escalates to human agents.

The customers quickly embraced the solution, expressing appreciation for instant, accurate information. “The digital assistant made the WhatsApp channel more accessible and appealing,” highlights Engin Kavas, Chief Technology Officer at Aydem Energy. “As a result, WhatsApp usage surged from 5% to 10%."

Lifting the call center burden

Today, the digital assistant manages about 1,000 customer inquiries daily through WhatsApp, which represents 10% of total customers calling customer service. “Of the inquiries that come through WhatsApp, 75% are fully resolved by the digital assistant without the need for live agent support,” notes Tosun. The team expects these figures to rise. “Over the summer, we may receive 100,000 calls a day, and AI is projected to handle 30% of those interactions,” Kavas adds.

The digital assistant excels at providing accurate and timely responses to customer inquiries. “Of the customers who use the digital assistant, 90% never call customer service again. This means that our agents need to deal with 900 less calls,” Tosun notes.

This has led to significant cost-savings across the company. Tasks that typically took 2,5 minutes on the phone can now be resolved in seconds through WhatsApp. “Now the Azure Open AI powered digital assistant handles the equivalent workload of 10 people. In peak periods, it can scale to handle the workload of 100 people.” adds Tosun.

Aydem Energy’s call center staff appreciate the support. "Agents can now focus on outbound calls for marketing or sales-related topics instead of customer inquiries," Tosun notes. "Our agents are also very happy because this is new technology, and we're one of the first to implement it on a WhatsApp channel. We've achieved something great with this project."

Implementing guardrails for responsible AI

To enhance security, Aydem Energy specified that only phone numbers operating from Türkiye can access the digital assistant functions; others are transferred to an agent. The company also monitors potential risks like AI discussing politics or religion. "We conducted penetration tests before launching the digital assistant to ensure it could handle potential vulnerabilities effectively," says Tosun. "For example, we ensured that the AI doesn’t reveal technical system details or sensitive CRM information. These precautions have been effective in mitigating risks.”

Reducing the operational cost of the digital assistant was another key focus. While initially, the AI assistant was permitted to use the entire internal knowledge base, it also resulted in significant data processing and higher costs. "We then created classifications by analyzing customer needs and scenario possibilities and limited the digital assistant to only access the library relevant for the specific scenario," Tosun explains. The company also addressed the risk of ‘token attacks’, or attempts to overload AI systems by submitting excessively long or complex inputs by directing long, out-of-scope or repetitive inquiries to live agents. “This approach reduced the cost per conversation by 75% compared to what it was when we did the initial POC," Tosun reports.

Exploring new opportunities

Building on this foundation, Aydem Energy is already planning further innovations. "We are working with Softtech to develop new functionalities for the digital assistant," says Tosun. "However, that's not our only focus. Another project we're planning with Softtech is an AI assistant to support call center staff directly." This assistant would provide instant support to agents during calls, eliminating the need to consult supervisors. Tosun sees great potential in these developments. "AI isn’t just helping us handle cases; it’s also creating opportunities for agents to reach new customers, increase revenue, and improve customer satisfaction rates. I can confidently say that we've proven that AI creates value. Moving forward, our future projects will follow the same principles, and we'll continue exploring new opportunities with Softtech and Microsoft," he affirms.