An engaging and interactive evening, brimming with pivotal discussions about the potential impact of Generative AI on customer interactions, among twelve seasoned digital leaders. This encapsulates the essence of the roundtable, organised by CIONET Belgium on Thursday, September 28, in alliance with our partner, Salesforce, and nestled within the elegant confines of restaurant “La Route” in Puurs.

An initial exploration into the mindset of the fellowship at the table taught us that AI has evolved beyond a mere buzzword into a technology that harbours the potential to redefine how businesses operate, deliver services, and forge value. The core of discussions among organisations is now transitioning from the feasibility of AI adoption to methodologies for its ethical, effective and secure implementation, aiming to strengthen business functionalities and enhance customer interactions and experiences.

Joachim De Rouck, Data & Analytics Manager at Brussels Airport Company, advocated that the starting point should invariably be the company’s strategic plan. For Brussels Airport, the optimisation of the passengers' journey, alongside the mitigation of CO2 emissions, stands paramount. Similarly, Frédéric Martin, Head of Automation & Innovation at Ethias, expressed, “My interest isn’t grounded in innovation or AI per se, but rather in procuring and implementing innovation where it significantly assists in actualising the company’s organisation or masterplan—whether through simplification or creating novel added value services for our clients. AI is simply another tool to employ. It necessitates evaluation based on its merits within the company’s specific context.”

Pieter Deurinck, VP, Head of IT Technology Strategy & Digital Business Transformation Solutions at UCB, disclosed that their involvement with AI has been intense and spans over a decade. However, today, they witness an explosive acceleration of possibilities with Generative AI. They partnered on enabling scientists to reason with AI in research and development of new patient solutions. Pieter envisions that all enterprise and other software vendors will soon undergo a quantum leap due to the integration of Generative AI into their solutions. Moreover, the executive leadership of the organisation decidedly exhibits a fervent interest in grasping this opportunity.

Regarding ChatGPT, while most participants permit its usage within their organisations, they impose clear guidelines to govern its application. Engie, for instance, has a governance board that restricts its use to meticulously defined use cases. Ethias, too, has released corporate guidelines on utilising ChatGPT. One prevalent risk identified is that internal, sensitive, or customer data might be directly or indirectly input into the tool, thereby exiting the company through that channel. Another risk is located in the domain of IP (intellectual property). There exists a substantial grey area, primarily concerning the type of data utilised to train and fuel the system. The ambiguity about whether personal or confidential data has been used by the tool, and the assurance of respecting data ownership, persists. Indeed, the outcome of the ongoing litigation between the New York Times and OpenAI over intellectual property rights could dramatically reshape the landscape concerning IP rights.

Numerous organisations are concurrently exploring how to implement AI internally, on their own data, and are keen on ensuring its use in a compliant and secure manner. Ethias, for instance, is orchestrating a secure environment for its internal use. A prudent approach, it was advised, could be to download a Llama language model, train it, and employ it strictly internally.

All participants echoed the sentiment of David Van Puyvelde - Senior Director, Solution Engineering Architecture Practice at Salesforce - that the unique context of each sector—including the sensitivity of the data handled, the degree of regulatory restrictions related to AI, security, or privacy, and the level of digitisation—vastly influences the velocity and impact of the application of (generative) AI in their business processes.

Linda Van Dijk, Director of Data Science & Governance at DKV Belgium, exemplified this by narrating her personal journey through two disparate periods in her career. The first period, exceeding 6 years, unfolded in Abu Dhabi at the company DAMAN, which was managed by the same mother company as DKV, namely Munich Re. In contrast, the circumstances today in Europe, where she currently works, differ drastically. The regulatory environment in Europe is substantially more rigorous, health-related information is managed with utmost confidentiality, and the company’s risk appetite is comparatively lower. Consequently, Linda is concurrently focusing on structuring information and data flows and overseeing the data governance process. The prime priority is on data defence, ensuring that all data is well-guarded and that data usage strictly aligns with all regulatory and legal constraints. Her approach remains pragmatic, implementing actions incrementally and exclusively employing models with the highest accuracy rates.

Geert Leekens, Enterprise Architect, and Joris Verberckmoes, Delivery & Operations Manager IS M&S Belux at ENGIE Electrabel, outlined how, via systematic exploration of data, they derived insights that facilitated the enhancement of customer lifetime value by presenting customers with innovative services and solutions. ENGIE and DKV, both amidst a transformational journey, are harnessing the plethora of accessible data, with AI poised as the quintessential tool to achieve this.

Intriguingly, the current use of AI has reached a juncture where it is altering the foundational elements of service provision and solution creation. At DEKA and ENGIE, the key metric is indeed the customer's lifetime value. Geert and Joris, who focus on ENGIE’s consumer side, initiated their journey with AI, primarily for energy consumption forecasting. Their mission evolved, and they have a history of deploying AI applications like chatbots, virtual assistants, and smart thermostats, culminating in applications for advising solar panel owners, enabling the provision of personalised recommendations for predictive maintenance. This application represents one of the initial instances where ENGIE pivoted towards offering long-term, value-added services to customers, enhancing their loyalty and optimising their lifetime value—a clear win-win for both entities. This strategy has morphed into a comprehensive service, aligning with their new strategy of ENGIE becoming a partner and coach to the customer in the context of the energy transition, guiding them on when to consume, produce, and store energy.

Naturally, discussions also ventured into the ethical dimensions of deploying AI, exploring how to balance data utilisation and advanced tools like Generative AI with customer trust and stringent adherence to data privacy norms. The focus then transitioned to the question: what about the impact of Generative AI on our jobs? Is there a concern about being superseded by AI engines? Most participants at the table, being tech optimists, believe AI can elevate our jobs, much like automation (RPA) has removed the monotonous and repetitive tasks from many jobs, AI can similarly impact our efficiency, albeit in a more advanced manner. “But this time”, commented Tom Asselman, Head of Enterprise Sales Belgium at Salesforce, “the logic diverges: it isn’t the foundation of human-centred or service-oriented jobs that are at stake, but rather the high-cost data processing jobs that could be pressured. The nurse may be less endangered than the doctor, the bookkeeper less so than the financial analyst. The pivotal question then becomes; how does society prepare for this future?”

Our sincere gratitude extends to Salesforce, Linda Van Dijk, Geert Leekens, Joris Verberckmoes, and all participating digital leaders for their invaluable insights into enhancing customer relations through AI, showcasing its expansive potential in optimising business operations, and illustrating the prospective future trajectory of the technology.

--