Exploring Cegeka's AI and Generative AI Prowess

At C-suite level, artificial intelligence (AI) and generative artificial intelligence (Gen AI) surged to the forefront of discussions, becoming an unavoidable topic in business circles. Whether the conversation concerns innovation, efficiency, productivity, cost-savings, or strategy, AI/Gen AI consistently commands attention and consideration.

While the buzz around Gen AI is considerable with PAC estimating Gen AI IT services CAGR at 74% in Belgium between 2024-2028, many discussions surrounding Gen AI are still centered around ideation and brainstorming, with both businesses and IT service providers actively identifying hundreds of potential use cases. However, significant challenges persist, particularly concerning data quality, accessibility, privacy, and risk management, as well as determining which specific use case(s) effectively address pressing business challenges.

IT services provider Cegeka, headquartered in Belgium, addresses these challenges through its Cegeka Data Solutions (Data & AI) team, consisting of over 400 specialists scattered across Europe (Belgium, Netherlands, Sweden, Italy, Romania, and Austria) with strong domain expertise in business intelligence, AI, data engineering, data governance & strategy, and Internet of Things (IoT).

For years, Cegeka has followed a reference architecture underpinned by Microsoft Azure and in-house data platforms for data science and data engineering projects. The firm has enriched and matured its data and AI models, enabling it to support its clients in four AI/Gen AI domains: Computer Vision, AI-infused Customer Analytics, Time Series Analysis (with IoT sensor data), and Natural Language Processing (NLP).

As an example, Cegeka used its cloud reference architecture (Microsoft Azure Cloud) to develop a new data platform for one of Belgium’s biggest utility providers, Fluvius. The data platform is capable of supporting the rollout of digital meters across Flanders, which aims to have smart meters installed in 80% of Flemish households by the end of 2024. With the ingestion of huge volumes of smart meter data every 15 minutes, Cegeka deployed AI to forecast the electricity consumption for the following day. The single integrated central data platform also facilitates the introduction of new services (e.g., energy sharing), identifying power consumption, better managing energy flows, and securely sharing data insights with multiple third parties.

According to Cegeka, customizability is a big requirement among its customer base as organizations do not solely want to talk to their data, but may also want to compare recent to historic data, among other use cases.

While custom AI/Gen AI projects are one of Cegeka’s strengths, underpinned by its reference architecture, the IT services provider is one of the few local champions that can productize AI/Gen AI with offerings, including Vida.Health, a platform-as-a-service (PaaS) for healthcare professionals using data and AI for greater insights; Mobilize, a platform comprised of AI traffic models for traffic intensity, flow, and management with simulation and forecasting capabilities; Capacity, a smart parking platform capable of smart parking allocation, dynamic capacity, real-time management and insights, and more; and Milo, an AI digital assistant and generative AI-accelerator with competencies in HR, research, mobility, municipality, service desk, and patents. The Milo framework delivers value of Gen AI in enterprise context fast and secure, and highly integrated in the existing business processes.

For example, Cegeka implemented a Gen AI digital assistant (Milo accelerator) at a group of local governments to give quick and accurate answers to a broad range of questions that inhabitants typically ask. For the municipalities a dashboard was created where they can meticulously analyze and evaluate citizen support conversations, considering various factors such as topic, sentiment, and satisfaction levels. This effort culminates in a comprehensive dashboard offering monitoring and insights capabilities to enhance citizen services, gauging success/failure rates, and expanding the knowledge base.

However, with Gen AI responses, there is always a degree of risk involved due to the quality of accessible data. To mitigate Gen AI risks, i.e., hallucinations, Cegeka quantifies the AI/Gen AI model performance based on its output, which is checked by an expert such as a service desk agent, healthcare clinician, etc. (human-in-the-loop) depending on the use case. Cegeka further emphasizes the importance of AI/Gen AI monitoring dashboards to track responses, sentiment, engagement, and customer satisfaction levels in order to see if the model needs to be adjusted to boost efficacy.

In PAC’s view, Cegeka is one of Belgium’s frontrunners in the AI/Gen AI domain, distinguished by its commitment to a data-driven and AI-centric strategy honed over recent years. This commitment is reflected in its roadmap, reference architecture, adaptability, and product development trajectory.

AI and Gen AI productization is a differentiator in the Belgian market and a path that other renowned Belgian IT services providers are yet to take, as navigating the complexities of commercializing a swiftly evolving technology entails inherent risks. Notably, niche technology start-ups and scale-ups possess the agility to outshine broader product offerings by prioritizing specialization in specific areas. For Cegeka, a combination of domain expertise, market awareness, a rapidly expanding Data & AI team, its cloud reference architecture, and plug-and-play data and AI layers provide a real advantage and make it possible to rapidly adjust product offerings according to evolving market and customer needs.