Recommendation tool
AI-powered tool that analyzes user needs and recommends the right product configuration. Reducing friction in product discovery and increasing user activation.
Timeline
3 weeks
My Role
Design & AI Integration
Team
Solo project
Tools
v0, OpenAI, React
Background
The suite covers audience intelligence for X and TikTok, digital intelligence for Meta, digital intelligence for LinkedIn, and social listening for X. New users often struggle to identify which tool best fits their specific use case.
The Insight
User research revealed 40% of trial users abandoned the platform without trying the right product for their needs. They were overwhelmed by options and unclear on capabilities.
The Opportunity
An AI assistant that understands user goals through natural conversation and maps them to the optimal product configuration—turning confusion into confidence.
Problem
Users don't know which product to use. Feature comparison tables are complex and sales calls create friction in the trial experience.
Goal
Create a self-service recommendation engine that guides users to the right product in under 2 minutes, increasing trial-to-paid conversion.
Solution
An AI-powered questionnaire that analyzes user needs (industry, goals, team size) and provides ranked recommendations with relevance scores and setup guidance.
How It Works
From user input to personalized recommendations in three steps
User Input Analysis
User describes their industry, target audience, and goals. AI parses intent and identifies key requirements from natural language.
AI Matching Engine
Capabilities are scored against user needs. Algorithm weighs feature relevance, use case fit, and team requirements to rank products.
Ranked Recommendations
Top 3 products displayed with relevance scores, matched capabilities, and ready-to-use queries to get started immediately.
Key Features
Designed to reduce friction and accelerate time-to-value
Relevance Scoring
Each recommendation shows a percentage match based on how well the product's capabilities align with stated user needs. "95% relevance" builds confidence.
Matched Capabilities
Visual tags show exactly which features match user requirements: "community management", "follower analytics", "engagement tracking" for quick scanning.
Ready-to-Use Queries
Pre-built queries tailored to user's industry and goals. Click "Access Tool" and start with relevant data immediately—no setup friction.
Setup Guidance
Step-by-step instructions for each recommended tool. Users know exactly what to do next, reducing support tickets and increasing activation.
Design Process
From user research to working prototype in 3 weeks
Week 1: Research & Strategy
- Analyzed trial abandonment data and support ticket patterns
- Interviewed 8 users who churned during trial to understand blockers
- Mapped the products and their capability matrices
- Defined the matching algorithm logic and scoring weights
Week 2: Design & Prototyping
- Designed conversation flow for natural language input
- Created recommendation card UI with relevance scores
- Built interactive prototype in v0 for user testing
- Iterated on copy to reduce cognitive load
Week 3: AI Integration & Launch
- Integrated OpenAI for natural language understanding
- Fine-tuned prompts for accurate capability matching
- Added analytics to track recommendation accuracy
- Deployed as widget in trial onboarding flow
Desireable Results & Impact
+35%
Trial activation rate
90s
Avg. time to recommendation
-28%
Support tickets
92%
Recommendation accuracy
Key Learnings
- Natural language input dramatically outperformed structured forms in user preference tests
- Showing relevance percentages increased user confidence and reduced second-guessing
- Ready-to-use queries were the #1 feature users mentioned in feedback surveys
Interested in this approach?
Let's discuss how AI-powered recommendations can improve your product onboarding
Ask for a demo