Small business owner using Business AI
← Home

TL;DR

Business AI on Messenger is an AI agent that helps small businesses respond to customers 24/7 — trained on their business context, up and running in minutes, and replies just like they would.

Role
Senior Product Designer
Timeline
April 2025 – May 2026
Team
I led a team of 3 designers on the self serve workstream, collaborating with PM, Eng, DS, and UXR
Goal
2025 → Prove product-business fit in our pilot program
2026 → 200k weekly active businesses

I led a team of 3 designers to build the 0→1 product for self serve businesses across web and mobile. I owned onboarding, activation, product upload and management, and the design of Business AI Home. In just 1 year we grew from a small pilot to over 1M weekly active SMBs.

Onboarding screen 1 Onboarding screen 2 Onboarding screen 3

Problem

Small and medium-sized businesses using Meta's self-serve tools are often overwhelmed by the volume of messages they receive across Meta's platforms. Without the staff or tools to keep up, conversations go unanswered — and customers move on.

Businesses overwhelmed by messages

Consumer experience

Consumer experience

Business experience

Solution

Business AI responds instantly to customers and helps with gathering customer information, driving sales, and delivering customer service by resolving common issues and escalating complex cases to human agents.

Process

We ran weekly research sessions with our Early Adopter Program and conducted three immersion trips to get direct feedback from small businesses in market. Working in 6-week sprint cycles, we prioritized shipping and iterating quickly over getting it perfect upfront.

Business AI immersion trip 1 Business AI immersion trip 2

Prototyping

I took an AI-first approach to prototyping — ranging from quick, scrappy concepts to high-fidelity builds, all made with Claude, Cursor, and VS Code. I also built out templates utilizing the design system, which helped me and the team move faster.

Integrating GDrive

Day 0 experience

Conversational agent

Iterations

I led the direction of onboarding and activation — getting this right was critical, as it was our customers' first impression of the product. Onboarding completion improved by +44% and activation by +57%.

V1

V2

V3

Improved Discoverability

NUX

The mobile NUX improved CTR by 6x

Optimized Onboarding

Loading

Iterations on the loading screen alone led to +12% onboarding completion rate

Enabled Product Upload

Price list

We saw +12% increase in retention when users complete this step

Business AI success

Results

  • In 2025 we proved out product market fit
  • In 2026 we increased from 10K → 1M WAB (weekly active businesses) over 6 months
  • 100x growth YTD in H1
  • 1M AI deep conversations / day

Other Work

Before joining the Business AI team, I led strategy and vision for Meta's business AI agents — defining AI foundations and patterns such as feedback mechanisms, generating states, and error handling. My primary focus was Meta AI Business Assistant, an AI agent embedded in Ads Manager to guide businesses through every step of the advertising journey.