How Incomplete
Business Information
Became AI Opportunity
Luiz Ottino
The story of integrating Gemini AI into Google for Retail and launching at Google Marketing Live 2023
In early 2023, Google faced a silent crisis that was costing millions in potential revenue: small business owners were abandoning their Google Business profiles halfway through setup, leaving critical information blank and hurting their visibility in search results.
As Google's AI capabilities evolved under the Bard initiative (now Gemini), our small team of four—a product owner, two engineers, and myself leading design—saw an opportunity to solve this problem using cutting-edge AI technology. We had just five months to deliver a solution for Google Marketing Live 2023.
The Problem: A 70% Abandonment Rate Nobody Talked About
The data was stark. Business owners would start creating their Google Business profiles but consistently abandoned the process when faced with lengthy forms asking for detailed descriptions, hours of operation, and business characteristics. The result? Incomplete profiles that performed poorly in search rankings, reducing both business visibility and Google's ad revenue potential.
Traditional solutions focused on simplifying forms or providing better guidance. We took a different approach: what if AI could do the heavy lifting instead?
The Solution: AI as a Business Partner, Not a Tool
Our Gemini-powered solution transformed the profile completion experience by:
Automatically suggesting profile content using Google's existing business intelligence—turning blank "About the Business" sections into compelling, contextually relevant descriptions
Generating smart recommendations that went beyond basic information to identify nuanced characteristics like "kids-friendly" environments or optimal operating hours based on industry patterns
Consolidating customer feedback from Google Business Reviews into actionable insights, helping owners identify trends like seasonal seating preferences or recurring service requests
Technical Innovation: The Progressive Disclosure Pipeline
The technical challenge was significant. Each AI request generated 15 response options, but showing all options would overwhelm users and increase costs dramatically. Our solution required building a sophisticated AI pipeline that could balance power, efficiency, and user experience.
The Gemini Knowledge Base Architecture
The pipeline drew from four critical data sources within the Gemini Knowledge Base to generate contextually relevant business content:
Business Category Classifications - Understanding industry-specific characteristics and common customer needs, from bakeries to photo labs.
Descriptions of Similar Businesses - Learning from successful business descriptions within the same category to understand effective positioning patterns.
Previously Inserted User Information - Incorporating existing data the business owner had provided, ensuring consistency with their established voice.
Customer Comments from Business Reviews - Analyzing actual customer feedback to identify authentic business strengths and unique characteristics
The Progressive Disclosure Flow
When a business owner requested AI assistance, the system would:
Process the comprehensive data foundation to understand the business context and customer sentiment
Generate 15 diverse content options, maintaining consistency with the business's tone
Apply intelligent filtering to surface the most relevant suggestions
Present through progressive disclosure - showing top 3 options initially, with pathways to "More than 15 items" or "New Tone of Voice" for complete regeneration
This created a feedback loop where user preferences refined the AI's understanding, while preventing decision paralysis and controlling API costs. The architecture proved that sophisticated AI integration doesn't require unlimited resources—it requires intelligent design.
Trust Through Transparency
The key to adoption wasn't just good AI - it was trustworthy AI.
We made the AI's reasoning process visible, showing users why certain recommendations were made and always maintaining their editorial control. Business owners weren't trusting a black box; they were collaborating with a transparent tool that enhanced their expertise.
Impact: More Than Just Better Profiles
The results exceeded expectations:
Dramatically reduced abandonment rates in profile completion
Enhanced data quality leading to improved search visibility
Increased engagement with Google's business management ecosystem
First-to-market positioning as the only major platform with conversational AI in business management workflows
Most importantly, we removed a significant friction point that was preventing SMBs from fully engaging with Google's advertising and visibility tools, creating a virtuous cycle of better profiles, better search performance, and higher platform value perception.
Three Key Lessons for AI Integration
AI succeeds when it reduces effort, not adds complexity. Focus on elimination and automation rather than feature addition.
Transparency and user control are prerequisites for AI trust. Making AI reasoning visible and keeping users in control of final decisions is essential for adoption.
Market timing matters tremendously in AI. Being first to market with a polished AI experience creates significant competitive advantage.
The Bigger Picture
This project didn't just solve a profile completion problem—it established a new paradigm for AI in business tools. By successfully integrating Gemini into a practical, cost-effective solution, we demonstrated that AI's greatest value comes from making complex tasks simple, not making simple tasks complex.
As we presented the feature at Google Marketing Live 2023, it became clear that we hadn't just built a feature—we'd created a foundation for the future of AI-assisted business management. The key insight remains: the most successful AI implementations are the ones users don't have to think about, they just benefit from.