CIBC Smart Search overview

Design Concept · CIBC Mobile Banking

Design Concept — not a shipped product

CIBC Smart Search

Reimagining mobile banking search as an AI-powered conversational guide, turning the app from a passive record of transactions into a tool that actively helps users understand their finances.

Timeline Oct 2025
Platform iOS
Role Product Designer
Methods Competitive Analysis, AI Interaction Design, Concept Prototyping

Key financial insights are buried in a long, scrolling feed

The CIBC app surfaces useful information like spending comparisons, budget alerts, and account summaries, but delivers it all in a chronological feed. Valuable insights get pushed down by low-priority notifications before users ever see them.

The question was how to make a user's financial data instantly accessible without rebuilding the whole app.

Current CIBC Advice tab with buried insights

Current Advice tab

Uncovering the "buried value" problem

Spending comparisons and budget alerts were lost in a chronological feed alongside low-priority notifications. Search was limited to articles, leaving users with no way to ask a direct question or surface a past insight. Valuable data was in the system, but there was no way to actively retrieve it.

CIBC Insights showing monthly spend data buried in the feed

The Insights tab showing a monthly spend comparison, one of many valuable data points buried deep in the feed

Market trends: AI-powered search

I looked at how leading companies like YouTube and Reddit are integrating generative AI. The best AI experiences are context-aware, guided, and action-oriented, surfacing what users need rather than waiting to be asked. That became the brief for Smart Search.

Market trends: AI-powered search in Grok, Reddit, and YouTube

AI-powered search across Grok, Reddit Answers, and YouTube, each pairing a natural language input with a guided, contextual response

Introducing Smart Search

Smart Search is a new AI-powered hub that sits on top of the existing Insights engine, turning the experience from a passive feed into an interactive guide. Users type a question in plain language and get back a structured answer drawn from their own data.

Every response is layered: a summary at the top, supporting data below, and a suggested action at the bottom, closing the loop between awareness and doing something about it.

Smart Search home with suggested prompts

Smart Search home with guided prompt suggestions

Smart Search conversational response with next action

A conversational response with supporting data and a suggested next step

Unlocking the value of what already exists

This project reinforced that great design is not always about adding something new. The data was already in the system. Introducing a conversational search layer shifts the app from a place where users passively receive information to one where they can actively seek it out and take control.

Next Project

CTV News: Local Video Discovery

Redesigning the Local tab to surface buried newscasts