Customer Ignite Talk: Ravneet Shah (CTO, Allica Bank) & OpenAI

Key Concepts

• Agentic Applications: AI systems capable of performing tasks, making decisions, and interacting with external systems (like email or portals) to achieve specific goals.
• Squadlets: A refined, smaller version of the traditional "Spotify model" squad, optimized for speed and cross-functional autonomy.
• T-Shaped Model: A professional development framework where employees possess deep specialization in one area while building adjacent skills to handle broader tasks.
• Deterministic vs. Non-deterministic Agents: A hybrid approach combining rule-based logic (deterministic) with generative AI (non-deterministic) to handle complex workflows.
• Relationship Banking: A business model prioritizing human-to-human interaction, which Allica Bank augments with AI rather than replacing it with chatbots.

---

1. Organizational Transformation and AI Adoption

Allica Bank, a UK-based SME bank, has transitioned from experimental AI usage in 2023 to a high-adoption model.

• Adoption Metrics: The bank moved from 25% organizational adoption to a median workday usage of 77%.
• Cultural Shift: Ravneet emphasized the need to "unlearn" traditional operational habits to embrace AI-native workflows across all departments, including finance, operations, and product engineering.
• Operating Model: The bank moved away from large, siloed teams to "squadlets." These are smaller, highly autonomous units that combine product, design, and engineering roles. The goal is to enable all team members—including designers and product owners—to ship code to production by the end of the year.

2. AI in Lending and Underwriting

Lending is the core business of Allica Bank, and the company has utilized AI to solve the "manual bottleneck" problem.

• The Challenge: Brokers and introducers often submit incomplete applications via email, which traditionally required manual intervention.
• The Solution: The bank deployed AI agents to ingest email communications, identify missing information, and proactively request data from brokers before the application enters the internal portal.
• Performance: This automation has significantly reduced the "time to decision," with targets reaching between 7 and 12 minutes for specific lending processes.

3. Augmenting Relationship Banking

Rather than replacing human relationship managers with chatbots, Allica Bank uses AI to provide "contextual intelligence."

• Strategy: AI is used to synthesize customer data and insights, allowing relationship managers to spend less time on administrative research and more time on high-quality, value-added interactions with clients.
• Philosophy: The bank maintains that technology should meet the customer where they are, rather than forcing customers to adapt to rigid, automated interfaces.

4. Engineering Velocity and Innovation

The bank’s focus on speed is evidenced by its deployment metrics and structural changes.

• Deployment Volume: The team achieved over 3,700 deployments in the previous year.
• Governance: By embedding all necessary decision-makers within the "squadlets," the bank avoids the "governance drag" that typically slows down regulated financial institutions. This allows for rapid iteration without needing to escalate decisions outside the immediate team.
• Future Goals: The objective for the next 12–24 months is to double the number of product increments—both customer-facing and internal (risk, compliance, security)—without sacrificing service quality.

5. Notable Quotes

• On Organizational Change: "There’s a lot that we need to unlearn before we start learning the new technology." — Ravneet, CTO at Allica Bank
• On AI Strategy: "We didn’t want to replace relationship banking by a chatbot... what we wanted to do was actually support them with the insights and the information that they need." — Ravneet, CTO at Allica Bank
• On Operational Philosophy: "We need to unlearn, but we also need to unstructure the way some of the teams have been structured to be able to rebuild them to move faster." — Clem, OpenAI

---

Synthesis and Conclusion

Allica Bank’s approach demonstrates that successful AI integration in a regulated industry requires more than just software implementation; it requires a fundamental restructuring of the organization. By shifting to a "squadlet" model, fostering a T-shaped skill set among employees, and using AI agents to handle complex, non-standardized inputs (like broker emails), the bank has successfully increased its deployment velocity while maintaining its core value proposition of relationship-based banking. The primary takeaway is that AI should be used to augment human expertise and streamline operational friction, rather than simply automating for the sake of efficiency.