�鶹������Ʒ

Taking the stage at Finovate 2026, �鶹������Ʒ Chief Revenue Officer opened with a direct challenge to financial institutions: the era of mobile banking primacy is ending, and the age of agentic banking has arrived.

The Three Phases of AI: How Fast We Got Here

West mapped out how rapidly artificial intelligence transformed since ChatGPT’s release in November 2022:

1. Phase One (The Brain): Centralized, conversational access to world knowledge.
2. Phase Two (Harnesses and Agents): A brief, 11-month transition where models gained memory, scaffolding, and digital “hands and eyes” through tools like open-source harnesses and Claude Co-work.
3. Phase Three (The Age of Context): Where we stand today.

“Think of a college student who just graduated and entered your company,” West explained. “They’re intelligent, but they don’t know what to do, how far to take it, or what ‘good’ looks like. You have to train them.”

For banks, investing in training AI with deep institutional and customer context is what turns raw model intelligence into reliable, compound value.

A Brief History of Banking Channels

To highlight the gravity of today’s shift, West placed agentic AI along banking’s 4,000-year timeline:

Just as smartphones displaced desktop-only banking, and as the internet displaced physical branch appointments before that, AI platforms are creating a brand new primary channel.

The Existential Threat: Consumers Aren’t Waiting for Banks

West explained that while most financial institutions spend their AI budgets on back-office automation and internal productivity, consumers are taking matters into their own hands.

Today, users are connecting their financial accounts directly into ChatGPT, MetaAI, Gemini, and others to analyze their spending and seek advice. This disintermediation poses a serious threat to traditional institutions. In response, banks are exploring three paths:

1. Consumer-Led AI Connection: Customers export their data directly into general-purpose LLMs (High risk of disintermediation and poor visibility for the bank).
2. Managed MCP/Connectors: Banks build controlled Model Context Protocol (MCP) integrations, allowing secure data flow and observability directly inside tools like Claude or ChatGPT.
3. Embedded Frontier: FIs bring native, frontier-level AI directly into their digital banking apps to offer deep, hyper-personalized advisory services.

Asking the Unanswerable Financial Questions

To prove why traditional digital banking falls short, West posed several common life questions consumers face daily:

• I’m buying a car. Should I pay cash, take out a loan, or lease?
• Will I be ready to pay for my kid’s college tuition when the time comes?
• Is now the right time for me to refinance my home loan?

Most banks lack the staff, tools, or real-time data integration to answer these questions at scale. Combined with open banking data and deterministic financial records, AI makes personalized advice accessible to every customer.

Seven Pillars of Agentic Banking

Building a trusted, actionable AI experience in a heavily regulated industry requires a deliberate stack:

1. Deterministic data: Your bank’s accurate, verifiable core records
2. Probabilistic AI: Generative models that bring conversational reasoning to the experience
3. Open Banking Data: Aggregated external accounts (mortgages, credit cards, investments) providing complete financial context
4. Auditability: Clear tracking of every decision and prompt output
5. Guardrails: Safety measurements ensuring calculations and advice remain accurate
6. Governance: Strict compliance controls suited for regulated environments
7. Actionability: Enabling the AI to safely move money, pay bills. Manage subscriptions, and update direct deposits on behalf of the user

How �鶹������Ʒ Helps You Own the Fourth Channel

Putting all seven pillars into production doesn’t require rebuilding your tech stack from scratch.

�鶹������Ʒ’s new Financial AI Assistant is powered by �鶹������Ʒ data and designed to drive growth within your platform. Combining the depth of financial intelligence and personalization with the natural flow of conversation, �鶹������Ʒ’s Financial AI Assistant delivers:

• Real-money answers in plain language: Consumers ask about their finances and get personalized guidance grounded in their real financial data.
• Proven, deterministic intelligence: Answers are computed directly from permissioned user data, ensuring every recommendation is accurate and reliable, never hallucinated.
• Full brand and voice control: Highly configurable guardrails keep your institution's unique voice, brand guidelines, and product recommendations in complete control.

By embedding frontier-level intelligence directly into your app, �鶹������Ʒ enables financial institutions to deliver proactive guidance, shift consumer behavior, and own the fourth channel.

What “Agentic” Really Means

Closing his keynote, West offered a clear definition:

“An LLM on its own is not an agent. Finance only becomes agentic when actions happen automatically and securely on the customer’s behalf with their control.”

When a paycheck hits an account, an agentic system automatically allocates savings, clears upcoming bills, and optimizes investments according to pre-set parameters. That is the future of automated finance, and the banks that build the infrastructure for it today will own the primary customer relationships of tomorrow.