Becco is a New York-based retail technology company that helps retailers measure and improve the performance of front-line sales associates. The platform collects real-time customer feedback at the point of sale through QR code surveys. These surveys capture star ratings, service attributes, and free-form comments about individual associates. The tracked attributes include whether an associate greeted the customer, listened to their needs, provided helpful support, demonstrated product knowledge, and offered relevant suggestions.
Problem: manual performance analysis limited AI coaching platform adoption
Retail managers receive large volumes of customer ratings, service responses, and written feedback. Although existing dashboards help them view this information, they do not explain how it should be used to coach individual associates. Managers must review survey responses, identify recurring performance patterns, interpret customer comments, and prepare coaching plans manually. This takes time, makes coaching quality dependent on manager experience, and makes it difficult to analyze comments at scale.
Performance gaps are not always identified consistently, strong service practices often go unrecognized, and associates do not receive structured weekly coaching. Becco found that many retail managers lacked the time or coaching expertise needed to translate customer feedback into specific development plans. The company therefore needed an AI coaching platform that could identify associate strengths, detect performance gaps, analyse customer sentiment, and generate practical coaching recommendations grounded in real customer feedback.
Solution: AI coaching platform for personalized associate development
GoML designed a secure AI coaching platform that converts Becco’s customer survey data into weekly associate performance insights and personalized coaching recommendations. The solution uses AI Matic’s Content Generation blueprint, supported by reusable ETL and data ingestion components to process survey, associate, and customer feedback data.
The platform connects with Becco’s existing AWS infrastructure, analyses service trends, evaluates customer comments, and generates manager-ready coaching plans. The Phase 1 MVP focuses on backend development and automated weekly processing, while Becco continues using its existing dashboard.
Secure survey data integration
The AI coaching platform connects to Becco’s AWS RDS database through secure, read-only access. This allows the system to retrieve the information required for performance analysis without changing Becco’s existing database structure or survey collection process.
The data pipeline processes:
• Customer star ratings
• Service attribute responses
• Free-form customer comments
• Associate profile information
• Store and regional hierarchy data
• Historical survey submissions
The system aggregates this information at the associate level and prepares it for AI-based analysis. Only associates who meet the agreed minimum survey threshold are included in the weekly coaching process.
AI-based associate performance analysis
The performance analysis engine evaluates each associate across the attributes tracked in Becco’s customer surveys.
These attributes include:
• Greeting customers
• Listening to customer requirements
• Providing helpful support
• Demonstrating product knowledge
• Making relevant suggestions
The system identifies repeated performance gaps instead of relying on individual survey responses. For example, it detects when an associate consistently receives lower scores for greeting customers or offering suggestions.
Customer comment sentiment analysis
Customer comments provide useful context beyond star ratings and service checkboxes. However, reviewing these comments manually becomes difficult as survey volumes increase.
The sentiment analysis module classifies each comment as:
• Positive
• Negative
• Neutral
These sentiment signals are combined with structured performance data to create a more complete view of associate performance.
Personalized coaching plan generation
The AI coaching platform generates personalized coaching plans for associates with recurring performance gaps.
Each coaching plan includes:
• The associate’s strongest service skills
• Areas requiring improvement
• Relevant customer feedback patterns
• Specific actions the manager should discuss
• Practical exercises or service improvements
• Areas to review during the next coaching cycle
For example, an associate might receive strong scores for product knowledge but lower scores for greeting customers and making suggestions. The system would recognize the associate’s product knowledge while recommending a consistent greeting approach, better discovery questions, and more relevant product suggestions.
Weekly automated coaching pipeline
The AI coaching platform runs through an automated weekly processing workflow.
Each weekly cycle:
• Retrieves survey and associate data from AWS RDS
• Groups responses by associate and store
• Checks whether associates meet the survey threshold
• Analyses performance strengths and gaps
• Classifies customer comment sentiment
• Generates personalized coaching plans
• Creates structured outputs for each store
• Stores results for use within Becco’s existing systems
The pipeline includes logging, error handling, and monitoring to support reliable scheduled processing.
Manager-ready performance outputs
The system creates structured outputs that store managers can use without reviewing raw survey records or rewriting AI-generated content.
Each associate-level output includes:
• Overall performance summary
• Strongest service attributes
• Recurring performance gaps
• Customer comment sentiment
• Positive performance recognition
• Recommended coaching actions
• Supporting evidence from survey patterns
This gives managers a clear starting point for weekly coaching conversations and reduces the time required to interpret customer feedback.
AI Matic development foundation
Becco required a new AI backend, survey data pipeline, sentiment analysis module, and automated coaching workflow. Since the Phase 1 project involved building a new AI system, GoML used AI Matic as the standardized development foundation.
AI Matic provided reusable engineering components, AWS infrastructure templates, model integrations, testing frameworks, and a consistent project structure. This allowed the development team to focus on Becco’s associate performance analysis, coaching logic, and manager-ready outputs instead of rebuilding common AI infrastructure.
The AI Matic foundation supported:
• AWS RDS data integration
• Amazon Bedrock and Claude Sonnet integration
• FastAPI and Python backend services
• Logging, monitoring, and error handling
• Testing of AI-generated coaching recommendations
Impact
- 85%+ accuracy in customer comment sentiment analysis
- 2x faster preparation of associate coaching plans
- 50% less manual review of survey data and comments
- 30% faster identification of performance gaps
- 2 to 3x more consistent weekly coaching across store teams
About
Before Gen AI and after Gen AI
“With an AI coaching platform, Becco transforms customer feedback into personalized weekly guidance that helps retail managers recognize strong performance and address gaps consistently.”
Prashanna Rao, Head of Engineering, GoML.
Key takeaways for retail technology platforms
Common challenges
- Performance dashboards do not provide coaching guidance
- Managers have different levels of coaching experience
- Customer comments are difficult to analyze at scale
- Generic coaching does not address individual behavioral gaps
- Associates with limited survey data require careful handling
Practical guidance
- Use survey data as the primary source of truth
- Set minimum data thresholds before generating coaching plans
- Combine structured feedback with customer comment sentiment
- Generate specific coaching actions instead of broad summaries
- Recognize positive behaviors alongside performance gaps
- Keep store managers involved in coaching delivery
- Start with scheduled batch processing before real-time coaching
Ready to build an AI coaching platform
Partner with GoML to build a secure AI coaching platform with reusable components, a standardized AWS architecture, and a production-ready development framework using AI Matic platform.
