Beyond the Hype: What GenAI Can’t Fix Without the Right Foundation
By Cindy Griffin, Financial Services Vertical Marketing Specialist at Smart Communications
The banking and financial services industry is at a crossroads with generative AI (GenAI). On one hand, customers are curious and generally optimistic about its potential. On the other, distrust of financial institutions, poor data quality, and fragmented regulations stand in the way of real and effective transformation.
The simple truth is this: AI can only be as effective as the foundation it’s built on. Let’s dive into why.
Trust and AI: The First Hurdle
The financial services industry is built on a foundation of trust. However, research from our 2025 Customer Experience and Communications Benchmark Report in Financial Services shows there is a trust gap. Globally, 37% of financial services customers don’t trust their financial institution—a number that rises to 38% in Singapore and 41% in Australia.
This raises a critical question: does introducing AI close or widen the trust gap? The answer depends not on the technology itself, but on how responsibly it is implemented. Transparent AI usage, clear disclosures, and reliable outputs can strengthen relationships and improve trust. Opaque applications and inconsistent results only widen the trust gap.
AI Strategy: Beyond the Wish List
Financial services customers are warming to responsible AI use as applications like multilingual translations and hyper-personalisation become more common. In fact, 61% of financial services customers believe AI/GenAI will improve their experience within the next five years. The appetite and the expectation that AI will be used in financial services is at hand.
If there were no regulatory or budget barriers, most financial institutions would leap at the chance to deploy AI across the customer journey. But each firm must face reality. Banking and financial services are highly regulated industries, and technology budgets are limited.
For example, would a firm archive every communication indefinitely if regulations didn’t require it? Probably not. However, adding AI to the mix makes the prospect of archiving even more complicated. The more data you have archived, the better your AI summarisations will be. Enterprise archive and retrieval typically assume a massive number of low-cost storage events and comparatively rare retrieval events at a much higher cost. As AI summarisations change your storage-to-retrieval ratio, your costs could explode. Without clean, structured data and thoughtful retention policies, even the best AI strategies become inefficient and expensive.
When the Foundation Is Weak
The risks of skipping foundational investments are significant:
- GenAI summarisation costs: A single workflow – retrieval, ranking, chunking, embedding, inference – multiplies costs compared to legacy search. At scale, expenses balloon quickly.
- Customer self-service: If AI draws on incorrect or conflicting data, customers get wrong answers. That drives repeat queries, higher service costs, and frustration.
- Employee experience: Staff relying on AI summaries of outdated or inconsistent knowledge bases end up wasting time untangling conflicts, often jumping between systems to find reliable information.
Without the foundation of clean data and robust governance for GenAI, the risks to the business are magnified.
Regulation: Fragmented but Inevitable
It’s important to understand the potential for future regulations around AI because the decisions made today around how it’s implemented often have implications 10 to 20 years from now.
Current global regulation of AI is fragmented, and financial institutions that compete globally must prepare for that reality. A few examples:
- Singapore: Prioritises transparency (explainable AI, disclosure) and data input quality. Uses principles-based frameworks aligned with OECD/G20 and co-develops standards with global tech firms.
- Australia: Uses a risk-based approach to encourage innovation while managing harm shifting toward mandatory guardrails. Ethics principles based on OECD norms and strong business input.
- New Zealand: Light-touch, principles-based guidance co-designed with stakeholders and aligned with OECD principles.
- EU: The EU AI Act is the most developed regulatory framework to date and has the most actionable insight into how jurisdictions may approach AI regulation. It classifies AI by risk level. High-risk use cases, such as those tied to sensitive communications or inbound requests, face strict obligations. Limited risk use cases, like correspondence generation, require transparency and disclosure.
Capturing Additional Information – Traceability & Content Provenance
The new era of AI calls for more than just having access to archived content. Firms must know:
- When and how AI influenced communications. Track AI usage across all processes (e.g., translations, tone changes, summarisations). This will help to comply with regulations that could become more stringent over time.
- The cost of large-scale summarisations. The storage-to-archive retrieval ratio is going to change significantly. Legacy search processes mean that archives are rarely retrieved. With GenAI summarisations, content will be retrieved at a much higher rate.
- The source and reliability of every data point. Know how, when, and by whom AI is used to trace the origin, history, and transformation of content. Content provenance is not optional; it’s foundational to trust and compliance.
Building a Smart Foundation for Effective AI
What will the future of customer communications look like with AI? Customer lifecycle data will evolve to create more types of communications in more steps of the customer journey feeding more systems. Every time something is archived, it will span multiple channels for GenAI to create more information and more content. And, the archive will not just be about outbound channels, but also an inbound source of information like a summarisation at contract renewal.
But this future of AI is not hypothetical for banking and financial services, it’s already here. Success depends on aligning a carefully designed strategy with the right foundation, enabled by a leading technology provider like Smart Communications:
- SmartIQ™ collects clean, structured data through guided digital journeys.
- SmartCOMM™ uses that data to create personalised, compliant customer communications.
- SmartHUB™ archives data in a way that supports efficient AI retrieval and summarisation.
- SmartPATH™ orchestrates intelligent, personalised customer journeys powered by trusted data.
Together, these solutions provide the backbone for successful AI adoption.
Realising AI’s Promise
GenAI can reshape customer experiences in banking and financial services, but only with the right foundation built on trust, strong data foundations, and responsible governance.
Financial institutions that invest now in systems to collect clean, structured data and digital archives to store that data will be better equipped to deliver effective AI and consistent omnichannel communications. Additionally, firms that focus on transparent AI usage and content provenance will find themselves more prepared for future AI regulations and frameworks.
AI can be a powerful tool to strengthen trust, or it can serve as a means to erode it. The difference comes down to the foundation you build. Build one that ensures transparency, and you’ll unlock AI’s true promise.