ING AI Chatbot: Building Smarter and Faster Banking Support

As digital interactions increase, banks are under pressure to deliver faster, more personalized support without compromising compliance, security, or customer trust. Traditional chatbots, built around scripted workflows and predefined responses, are increasingly struggling to keep up with the complexity of real customer conversations. This created the backdrop for the ING AI chatbot initiative.

The rise of generative AI has changed the conversation, not because it magically solves customer service, but because it forces banks to rethink what support systems should actually do. In an industry shaped by regulation, trust, and operational complexity, deploying conversational AI is as much a governance challenge as a technology one.

The ING AI chatbot has emerged within this broader shift.

Developed by ING in collaboration with McKinsey & Company and QuantumBlack, the initiative became an early example of generative AI entering customer-facing banking environments in Europe.

What began as a focused pilot soon became a broader case study in how enterprises can responsibly deploy generative AI in customer-facing environments.

The challenge: Managing high volumes of customer queries

Every week, ING receives approximately 85,000 customer interactions through phone calls and online chat in the Netherlands, one of its largest markets.

While the bank’s traditional chatbot could resolve roughly 40 – 45 percent of customer queries, a significant number of customers still required assistance from live agents. This meant nearly 16,500 customers per week had to wait for human support, often during limited working hours for non-urgent requests.

The challenge for ING was clear:

• Improve customer response times
• Reduce dependency on live support agents
• Deliver more personalized assistance
• Maintain strict banking security and compliance standards

Generative AI presented an opportunity to rethink the customer support experience from the ground up.

Building a next-generation AI chatbot: A seven-week collaboration

ING and McKinsey together assembled a joint team that worked intensively over seven weeks to design and deploy a generative AI-powered chatbot – capable of delivering faster, more personalized customer assistance while maintaining strict banking safeguards.

The project began with a deep assessment of ING’s existing chatbot infrastructure to identify pain points and gaps in customer experience. Based on these findings, the teams developed a sophisticated multi-step response pipeline to generate more accurate, context-aware answers.

The solution incorporated:

• Knowledge retrieval from multiple internal data sources
• AI-driven ranking of potential responses based on relevance and usefulness
• ‘Disambiguation’ capabilities, where the chatbot presents multiple possible answers if customer’s intent is unclear
• Real-time conversational assistance tailored to customer needs

Unlike traditional rule-based chatbots, the generative AI system was able to engage with natural language in a far more human and dynamic manner.

Building responsible AI guardrails

Because banking is a highly regulated industry, ING prioritized security, compliance, and responsible AI deployment from the very beginning.

Risk and compliance stakeholders were involved throughout the development process to ensure the chatbot operated within strict safety boundaries. Custom guardrails were built using tools from QuantumBlack Labs to prevent the chatbot from providing sensitive financial guidance, such as mortgage advice and investment recommendations.

“By prioritizing real customer needs, security, and ease of use, ING was able to develop a bespoke customer support tool that gives users the best possible experience”, says, Marnix van Stiphout, ING Chief Operating Officer

Testing the chatbot with real customers

The generative AI chatbot was initially rolled out as a pilot to 10 percent of real ING customers in the Netherlands using the bank’s mobile support chat function.

The results were significant. Customers experienced:

• Faster query resolution
• More detailed and personalized responses
• Reduced dependency on live support agents
• Improved overall customer satisfaction

And since its launch in September 2023, thousands of customers have interacted with the new-gen AI chatbot, making it the first-of-its-kind real-life customer-facing pilot in Europe.

The business impact of the ING AI chatbot

The speed with which the chatbot was built and deployed far outpaced the timelines required to develop previous industry-standard chatbots, which can take several years of programming and fine-tuning to get right.

What’s more? Within the first seven weeks of use, the new chatbot was offering a better customer experience:

It was helping 20 percent more customers resolve issues – compared to the previous solution, reducing pressure on customer service call centers and creating opportunities for more customers to shift from phone support to digital chat channels.

“When done right, using gen AI can be incredibly powerful in creating a better customer experience while also prioritizing the security of banking customers”, says Stephanie Hauser, McKinsey senior partner and co-leader of the Global Banking and Securities Practice.

Creating a foundation for enterprise-wide AI adoption

What began as a customer service enhancement initiative quickly evolved into a broader enterprise-wide AI transformation strategy for ING. Rather than treating the chatbot as a standalone experiment, the bank used the project as a foundation for scaling generative AI capabilities across the organization.

To support this vision, ING invested heavily in capability building across more than 50 support functions, including teams within Risk, Contact Center Operations, Analytics, and Technology.

Likewise, the collaboration between ING and the McKinsey team has created a path to double the chatbot’s performance in the near future. The aim is to build a scalable model that can be extended to all other ING countries.

“This project has helped establish a solid technical foundation that puts ING at the forefront of gen AI applications within the banking industry.”, says Bahadir Yilmaz, ING Chief Analytics Officer.

“Gen AI technology will further evolve and we feel well-positioned to leverage those developments in order to offer the customer the best experience and remain a technology leader in our industry.”

Key takeaways

Here are a few takeaways from this ING AI chatbot case study.

Core modernization in banks and financial institutions: A strategic imperative

Legacy systems hinder the adoption of new technologies due to high maintenance costs and a complicated, isolated design. Modernization offers a workable long-term strategy by delivering adaptable, secure, and user-friendly platforms which can satisfy the regulatory requirements and enhance customer satisfaction. Therefore, banks must upgrade or replace legacy systems through core modernization to remain flexible and competitive in an evolving digital banking landscape.

Data and knowledge integration are critical for AI success

The chatbot’s effectiveness depended heavily on its ability to retrieve and synthesize information from internal knowledge systems, highlighting the importance of strong data infrastructure in enterprise AI deployment

AI works best when combined with human oversight

The case study underscores the importance of human judgment. Rather than replacing human agents entirely, ING used AI to handle routine and repetitive queries, allowing support teams to focus on more complex customer issues that require empathy and judgment.  

AI adoption requires cross-functional collaboration

The project demonstrated that successful AI deployment is not just a technology initiative. Risk, compliance, analytics, operations, and customer service teams all played a critical role in building a safe and scalable solution.

Real-world testing is essential before full-scale rollout

By initially exposing the AI chatbot to only 10 percent of customers, ING was able to safely validate performance, customer response, and risk management before wider deployment.

Organizational capability building is just as important as technology

ING invested in training and capability development across more than 50 support functions, reinforcing the idea that enterprise AI transformation requires people readiness alongside technological readiness.

Pilot projects can become enterprise blueprints

The initiative demonstrated how a well-executed pilot project can move beyond experimentation to become a strategic blueprint for enterprise-wide AI adoption, helping organizations accelerate future innovation at scale.

In brief

For years, financial institutions relied on traditional chatbots and automated systems to handle basic customer inquiries and routine support requests. However, the rise of generative AI has significantly expanded these capabilities, enabling chatbots to understand context, engage in more natural conversations, and handle far more sophisticated and personalized customer interactions.