Techvicast Episode 4: Making AI Work In Fintech
Loading...
Marketing Manager - Techvify
Dive into TechviCast Episode 04, where host Vincent and guest Sachin Rajat Sharma share practical strategies to bring AI from concept to reliable production value in fintech and digital banking. Drawing on deep frontline experience, they show how modern composable platforms help institutions overcome legacy challenges, accelerate launches, achieve substantial cost reductions, and protect against emerging threats, all with innovation that remains secure and compliant.
Discover actionable insights: launch a digital bank in as little as 6 months, deliver major operational efficiencies (reduced engineering needs, automated marketing, rapid UX redesigns), unlock emerging revenue opportunities through AI-driven personalization, implement gold-standard fraud protections including advanced liveness detection and mule screening, and understand why establishing a modular foundation first is the key to fully realizing AI’s potential. Essential listening for fintech professionals eager to lead in 2026 and beyond.
Dive deeper into the conversation with this super-granular timeline! Vincent and Sachin take you on a lively, no-nonsense ride through the real-world world of AI in fintech. From tackling those pesky legacy headaches and smart strategic moves, all the way to hands-on AI wins, rock-solid risk defenses, and clear next steps to keep you ahead of the curve. Pick any section that sparks your interest and jump right in!
0:00 - 1:34: Introduction & Guest Welcome
Vincent introduces TechviCast, Techvify's AI focus, and welcomes Sachin, highlighting his 25+ years in banking from traditional to fintech.
1:34 - 2:57: Sachin's Career Journey in Banking Evolution
Sachin shares his experience from early SMS banking to today's AI/embedded finance era; banks seen as slow but innovative in fraud/customer service.
2:57 - 4:54: Is Banking Infrastructure Truly Fragmented?
Discussion on APIs, embedded finance, regulations, viewed as evolution/coexistence of old + new tech/models, not true fragmentation.
4:54 - 8:04: Strategic Advice for Incumbents vs. New Entrants
"Speedboats" for large banks to test innovations safely; new players: go AI-first without legacy baggage.
8:04 - 11:16: Legacy Renovation Risks & "Build New" Analogy
50-year-old building metaphor: renovate old core at high risk; build modern adjacent structures first, integrate gradually.
11:16 - 13:39: 101 Digital's Composable Platform Approach
Modular "digital bank in a box" for incumbents (add BaaS/embedded finance) and new banks (launch in 6 months vs. 18–24).
13:39 - 15:18: Modularity, Deployment Flexibility & AI Fit
Cloud/on-prem options; composable microservices enable better data access/logging for AI compared to monoliths.
15:18 - 17:46: AI in Legacy vs. Modern Architectures
AI thrives in modern stacks; legacy can adopt selectively (e.g., internal tools) but modern unlocks full potential.
17:46 - 19:50: AI Value Across the Fintech Value Chain
Onboarding (liveness/fraud), underwriting, servicing, compliance, backend ops, AI already embedded in many areas.
19:50 - 23:02: Safe AI Adoption in Legacy Environments
Start with internal efficiency (dev co-pilots, document parsing, chatbots for agents) before customer-facing risks.
23:02 - 26:21: AI Accelerating Development & Operations
Reduces engineering needs, enables rapid changes (e.g., 101 Digital's "Simpler" tool); balances speed with regulatory scrutiny.
26:21 - 35:40: AI: Cost Efficiency vs. Revenue Driver
Clear ops savings (fewer staff, faster dev, marketing automation); revenue potential emerging but early-stage, prioritize easy wins first.
35:40 - 40:30: Addressing AI-Heightened Fraud & Risk
Gold-standard controls (gov registries, advanced liveness/NFC, mule detection); upgrade defenses as threats evolve.
40:30 - 42:46: Where Do AI Failures Really Come From?
Often governance/platform/execution issues, not just models, need testing, policies, gradual rollout.
42:46 - 47:53: Prioritization: Platform First, Then AI + Closing Advice
Build composable foundation → layer AI; start safe/low-risk use cases to build momentum; adopt now or risk falling behind. Thanks and wrap-up.