SupTech has matured from isolated pilots to a strategic institutional capability, but scaling it sustainably requires more than good models — it demands strong data foundations, modular architecture, governance by design, and genuine human capacity. Reflecting on SupTech Week 2025, this blog argues that the next frontier is building systems that remain robust under future technological shifts, including the looming challenge of quantum readiness.
Author: Tatia Tsiklauri, AI & Data Science Specialist, National Bank of Georgia
This article represents the author’s personal perspectives based on attending SupTech Week 2025.
_________________________
Introduction
Over the past few years, SupTech has evolved from a largely experimental concept into a strategic capability for financial authorities worldwide. Artificial intelligence, advanced analytics, and data engineering are no longer peripheral tools, they are increasingly embedded in the core of supervisory, regulatory, and policy functions.
My own professional journey sits at this intersection, working on AI and SupTech initiatives within a central bank context, while actively engaging with international peers, innovators, and regulators. Events such as SupTech Week 2025 provided a valuable opportunity not only to exchange experiences, but also to reflect more critically on where SupTech is delivering real value, where expectations may be running ahead of reality, and what should be top of mind as we move into the next phase of adoption.
This blog brings together personal reflections from that journey, key takeaways from SupTech Week, and broader observations on how supervisory authorities can move from isolated AI use cases toward sustainable, institution-wide SupTech capabilities.
From Tools to Systems: Lessons from a SupTech Journey
Early SupTech efforts in many institutions, including our own, often start with narrowly scoped tools, a chatbot for internal regulations, a document classifier, a basic anomaly detection model, or a pilot dashboard. These projects are valuable, but over time a clear lesson emerges: isolated tools do not scale impact on their own.
From a SupTech perspective, the most immediate implication of quantum technologies relates to cryptographic resilience. Many supervisory systems, data exchanges, and reporting infrastructures rely on public-key cryptography that could become vulnerable in a future quantum-enabled environment. As a result, quantum readiness increasingly intersects with SupTech through secure data pipelines, supervisory platforms, and API-based supervision. The key lesson for SupTech today is not early adoption, but early alignment. Supervisory authorities should ensure that systems being built now are modular, crypto-agile, and adaptable to post-quantum standards, avoiding future lock-in to vulnerable technologies.
In this sense, quantum computing reinforces a broader message: SupTech is not only about adopting today’s tools, but about designing supervisory systems that remain robust under tomorrow’s technological shifts.
The real shift happens when institutions begin to think in terms of systems, not solutions. This requires re-framing SupTech not as a collection of AI models, but as an integrated capability built on four pillars:
- Data foundations – reliable, well-governed, and reusable data infrastructure
- Modular technology architecture – reusable services, APIs, and microservices rather than one-off applications
- Governance and risk management – model validation, explainability, security, and accountability embedded by design
- Human capacity – supervisors, analysts, and policy teams who understand how to work with AI, not just consume its outputs
Without progress across all four, SupTech initiatives risk remaining experimental, fragile, or underutilized.
Key Takeaways from SupTech Week 2025
SupTech Week reinforced several important themes that resonated strongly across jurisdictions, regardless of size or maturity level.
1. AI is Moving Closer to Core Supervision
A recurring message was that AI is no longer confined to innovation labs. Authorities are increasingly embedding AI into day-to-day supervisory processes, including risk prioritization, thematic reviews, entity profiling, and market monitoring.
However, this transition is gradual and deliberate. Institutions that are succeeding tend to start with decision support, not decision replacement, ensuring that human judgment remains central while AI augments scale, speed, and consistency.
2. Governance is the Real Differentiator
The most advanced use cases were not necessarily the most complex models, but those supported by strong governance frameworks. Topics such as model risk management, auditability, explainability, and lifecycle controls featured prominently in discussions.This reinforces an important point: in a supervisory context, trust in the system matters as much as technical performance. AI that cannot be explained, validated, or governed will not be used, regardless of its theoretical accuracy.
3. Platforms Matter More Than Pilots
Another consistent insight was the growing shift toward SupTech platforms. Rather than developing standalone tools, authorities are investing in shared platforms that host multiple AI services, data pipelines, and user interfaces under a common governance and security model.
This platform approach enables reuse, reduces duplication, and significantly lowers the marginal cost of introducing new use cases.
From 2025 into 2026: What Should Be Top of Mind
Looking ahead, several priorities stand out for SupTech leaders and practitioners.
1. Data Architecture as a Strategic Asset
AI initiatives cannot outpace data maturity indefinitely. Institutions should treat data warehouses, data lakes, and metadata management not as IT projects, but as strategic supervisory infrastructure.Well-designed data architecture enables faster onboarding of new use cases, more robust analytics, and better cross-departmental collaboration.
2. Designing for Longevity, Not Demos
There is increasing recognition that SupTech success should be measured not by pilots delivered, but by systems sustained. This means designing solutions that are maintainable, adaptable, and resilient to staff turnover, regulatory change, and evolving risks.
Key questions to ask early include:
- Who owns this system long-term?
- How will it be updated, monitored, and audited?
- How does it integrate with existing supervisory workflows?
3. Human-Centric AI in Supervision
SupTech is ultimately about empowering supervisors, not replacing them. Training, change management, and clear communication about what AI does and does not do are critical. Institutions that invest in AI literacy for non-technical staff tend to see higher adoption, better feedback loops, and more realistic expectations.
SupTech is entering a more mature phase. The question is no longer whether AI can support supervision, but how institutions can deploy it responsibly, sustainably, and at scale.The conversations at SupTech Week, combined with practical experience on the ground, point toward a clear direction: platform-based architectures, strong governance, robust data foundations, and human-centric design will define the next generation of SupTech.
As we move further into 2026, the most successful authorities will be those that view SupTech not as a collection of innovative tools, but as a core institutional capability aligned with their supervisory mandate.
_______________________________________________
For further information, we encourage you to read the State of SupTech Report 2025, access session recordings and engage in discussions on GovSpace.io.
