Technology advances fast, but meaningful innovation still comes down to one truth: systems work best when they’re built around people.

In risk, fraud, and compliance (RFC), this means designing tools that understand intent, reduce friction for investigators, protect sensitive data, and adapt as fraud evolves. The Grace Hopper Celebration (GHC) surfaced ideas that map directly to those challenges – not as inspiration, but as working principles for how we build next-generation financial crime analytics.

Below are the takeaways that matter most for practitioners.

Non-obvious thinking for emerging fraud thoughts

One of the most memorable moments came from Rohit Bhargava’s SIFT framework – space, insight, focus, twist. In other words, create space, observe differently, find focus and add a twist.

Bhargava, author and founder of Non-Obvious Company, left this message: “People who understand people always win.” I think this perfectly encapsulates the challenge of designing financial crime systems. It’s especially relevant in financial crime, where threats change faster than traditional systems can keep up.

What this means for RFC teams:

• Fraudsters constantly innovate, so our detection strategies must evolve just as quickly.
• Risk analysts notice patterns that models can’t yet quantify, but models can also often discover patterns that risk analysts overlook. This requires collaboration between tools and their operators to amplify their intuition or minimize the effects of unintentional bias.
• Non-obvious thinking is essential for addressing emerging threats, such as synthetic identities, deepfake-enabled social engineering, and instant-payment fraud.

In other words, curiosity is a risk-mitigation strategy.

Where product meets engineering

Utsha Sinha, a product manager at Uber, underscored something every successful fraud or AML system already proves true: strong products rarely emerge from silos. For banks and regulators, the bar is high – explainable models, auditable decisions, transparent data lineage and workflows that fit the realities of investigation work.

What this looks like in practice across industries:

• Explainable models regulators trust.
• Scenario engines that scale with global portfolios.
• Transparent lineage for data and decisions.
• Fraud detection that minimizes false positives.
• Workflows that fit real needs.

Delivering these outcomes requires tight alignment between PMs, engineers, architects, risk SMEs, data scientists and others. Fraud and AML tools only work as well as the teams building them understand one another.

As Sinha put it, “great products emerge not from silos, but from intentional cross-functional learning.”

Sustainable engineering as a core requirement

Tanaya Salunkhe, a software developer at Amazon, highlighted sustainability as a technical challenge, not a moral accessory. For real-time fraud scoring and large-scale AML workloads, efficiency is directly tied to performance, cost and compliance.

How this shows up in financial crime systems:

• Efficient model retraining helps banks reduce cloud spend while maintaining model freshness.
• Predictive auto-scaling ensures fraud scoring systems can handle peak transaction volumes without over-provisioning.
• Lightweight model architectures reduce carbon footprints and enhance deployment flexibility for global banks with stringent data residency requirements.

I believe that sustainable engineering is no longer something that is “nice to have”; it’s fundamental to the future of real-time financial crime analytics.

Cybersecurity’s human element

Neelima Kumar, an engineering manager at Uber, reframed security as a deeply human discipline. For banks and regulatory agencies, this couldn’t be more relevant.

Kuman said by implementing these practices in her work, she’s seeing:

• Machine-based policy approvals to reduce insider threat risk.
• Automated data minimization and anonymization.
• “Security scorecards” that encourage employee learning.

These principles align closely with SAS’ approach to data privacy, compliance, and model governance.

With this, there are some implications Kumar shared for RFC teams:

• Fraud and AML systems must protect sensitive financial data while enabling investigative efficiency.
• Model governance must be transparent enough for auditors and regulators to have confidence in.
• Investigators are more effective when security practices support, rather than hinder, their work.

Cybersecurity in the financial sector is about more than just threat prevention; it’s about empowering people to responsibly use sensitive data without unnecessary friction.

Expanding who gets to innovate

In her closing remarks, astronaut Kellie Gerardi said: “Progress isn’t just about exploring new frontiers – it’s about expanding who gets to explore them.”

That mindset defines both GHC and SAS’ mission in financial crime. We build systems that help global institutions reduce risk, fight fraud, protect communities and uphold trust. The work we do impacts millions of people worldwide, and it’s most impactful when grounded in empathy, human behavior, ethical AI, and sustainable engineering.

As I continue my journey in risk, fraud and compliance, I’m inspired by the women technologists leading these conversations and grateful for the chance to bring these insights into our work at SAS.

Together, we can continue building technology that protects the world’s financial systems and the people who rely on them.

Learn more about fraud and financial crimes solutions from SAS