Upcoming Event: The AI Agent Implementation Framework
Upcoming Event: The AI Agent Implementation Framework
Digital Sportsman
Tyner, North Carolina
Outdoor Recreation Technology (SaaS)
Amazon Bedrock, Amazon ECS (Fargate), Amazon RDS (PostgreSQL), AWS Secrets Manager, Amazon ECR (Elastic Container Registry), Amazon CloudWatch, AWS Certificate Manager, Application Load Balancer (ALB), Amazon VPC, GitHub Actions (CI/CD)
Digital Sportsman is an AI-powered SaaS platform that provides booking, CRM, marketing, and reporting tools to professional fishing guides, charter captains, hunting guides, and outdoor experience operators. The company needed a conversational AI assistant that would allow its professional users to manage bookings, look up client information, and query business metrics through natural language, replacing complex dashboard navigation with simple conversational interactions. Avahi delivered the VA for Pros Chatbot, built on AWS services including Amazon Bedrock, Amazon ECS, and Amazon RDS, featuring a dynamic NL-to-SQL engine that adapts to database changes without code modifications. The solution streamlined guide workflows, enforced zero-trust data security, and established an extensible AI framework for Digital Sportsman’s continued platform evolution.
Digital Sportsman is a technology company headquartered in Tyner, North Carolina, that provides an all-in-one booking and management platform purpose-built for outdoor experience operators. The platform serves professional fishing guides, charter captains, hunting guides, rental operators, tour operators, marinas, and retailers across the outdoor recreation industry. Digital Sportsman combines booking management, CRM, email marketing, resource and equipment tracking, reporting, and AI automation into a single platform, helping operators streamline daily operations, manage complex schedules, and drive repeat bookings. The company operates on a no-commission model and was recognized with a Bronze Stevie Award in the 2025 American Business Awards for its innovation in empowering outdoor professionals through technology.
Digital Sportsman’s professional guide users relied on complex dashboard interfaces to manage bookings, look up client information, and access business metrics. This manual approach required guides to navigate multiple screens and forms to complete routine tasks. Creating bookings involved filling out multi-step forms, retrieving client history meant searching through data tables, and checking revenue figures required navigating to specific reporting dashboards.
For professional guides whose primary focus is delivering outdoor experiences to their clients, this operational overhead directly detracted from productive time. Without an intuitive way to interact with the platform conversationally, guides spent valuable time on data entry and lookup tasks that could be streamlined through natural language. The challenge was compounded by the complexity of the underlying data model, a production PostgreSQL database spanning 50+ tables, which made it impractical for guides to self-serve complex queries about their bookings, clients, and financial performance.
If left unaddressed, this friction risked lower platform adoption among professional users and limited Digital Sportsman’s ability to differentiate its product in a competitive market for outdoor business management tools.
Digital Sportsman’s production infrastructure was already built on AWS, with its core PostgreSQL database hosted on Amazon RDS and application workloads running within an AWS VPC. This existing AWS foundation made it natural to extend the platform with AWS-native AI and compute services rather than introducing a separate cloud provider.
Amazon Bedrock provided managed access to foundation models, specifically Claude Sonnet 4 for natural language understanding and SQL generation, and Amazon Nova Premier for agent orchestration and tool execution, without requiring Digital Sportsman to manage model infrastructure. Amazon ECS with Fargate enabled containerized deployment of the virtual assistant backend with automatic scaling and no server management overhead. AWS Secrets Manager, Amazon CloudWatch, and AWS Certificate Manager provided the security, observability, and certificate management needed for a production-grade deployment integrated with Digital Sportsman’s existing environment.
Digital Sportsman engaged Avahi through an AWS referral, selecting the firm based on its credentials as an AWS consulting partner with deep experience delivering cloud-native solutions on the AWS platform. Avahi’s specialization in Generative AI was a key differentiator. The VA for Pros project required expertise in foundation model integration, prompt engineering, NL-to-SQL implementation, and agentic AI architecture: capabilities that sit at the intersection of cloud infrastructure and applied AI.
Avahi’s track record of building production-grade GenAI solutions on AWS services, including Amazon Bedrock, Amazon ECS, and Amazon RDS, gave Digital Sportsman confidence that the team could deliver a secure, scalable virtual assistant integrated with their existing AWS infrastructure. The combination of AWS consulting partner experience and hands-on GenAI specialization positioned Avahi to move quickly from architecture to production deployment.
Avahi delivered the VA for Pros Chatbot, an AI-powered conversational interface designed specifically for professional hunting and fishing guides. The chatbot enables guides to create bookings, look up client information, and query business metrics through natural language, replacing the need to navigate complex dashboard menus with simple conversational interactions like “What’s my revenue this month?” or “Show me bookings for John Smith.”
The solution architecture is built on a multi-agent framework powered by the Model Context Protocol (MCP). When a guide submits a natural language query, the system routes it through Claude Sonnet 4 via Amazon Bedrock, which classifies the query as either an informational inquiry or an actionable request. Informational queries trigger the NL-to-SQL engine, which passes the database schema and field value mappings to the LLM at query time, generating accurate PostgreSQL queries against the production RDS database. Action queries, such as booking creation, are routed to Amazon Nova Premier, which orchestrates MCP tool calls to execute the appropriate backend API operations.
The NL-to-SQL engine represents a key technical innovation. Rather than relying on pre-trained models or hardcoded query templates, the system dynamically injects the full database schema and enumeration mappings into the LLM context at query time. This approach means the chatbot adapts to database changes, including new tables, modified fields, and updated enum values, by updating schema files alone, with no code changes required. Business rule safeguards are embedded directly into prompts with conditional logic for known edge cases, ensuring data integrity across a complex 50+ table production schema.
Security was a critical design priority. User identity is extracted directly from the bearer token in each request rather than from the request body, eliminating the risk of client-side manipulation and enforcing a zero-trust approach to data access. Guide-specific data isolation is enforced at the query generation layer through automatic guide_id filtering, ensuring each professional only accesses their own data. Revenue calculations include deleted_at IS NULL filters to exclude canceled trips and return accurate financial figures. CORS whitelisting limits access to authorized Digital Sportsman domains, and JWT tokens stored in AWS Secrets Manager authenticate all API requests.
The backend runs as a containerized service on Amazon ECS with Fargate, with a CI/CD pipeline built on GitHub Actions that automates Docker image builds, pushes to Amazon ECR, and deploys updated ECS task definitions. Amazon CloudWatch provides logging, health monitoring, and alerting across the infrastructure. Conversational context is maintained in a dedicated PostgreSQL table on Amazon RDS, preserving session history across interactions for contextual continuity.
During implementation, Avahi identified and resolved a critical chatbot tone issue within 24 hours. The VA for Pros initially responded with customer-facing language that was not aligned for a professional guide audience, being verbose and unfocused. Through rapid prompt tuning, the team transformed the chatbot’s communication style to be professional, concise, and tailored to the outdoor industry context, with follow-up clarification questions for ambiguous requests such as “For what time period?”
The VA for Pros Chatbot transformed how Digital Sportsman’s professional guides interact with the platform. Guides can now create bookings, retrieve client information, and access business insights through natural language conversations instead of navigating multi-step dashboard workflows. The chatbot handles ambiguous requests with intelligent follow-up questions and returns guide-specific, accurately filtered results from the production database. Tasks that previously required navigating multiple screens and forms, such as creating a booking or checking monthly revenue, are now completed through a single conversational exchange.
The solution’s adaptive architecture positions Digital Sportsman for long-term scalability. The NL-to-SQL engine accepts updated schema files without code changes, allowing the chatbot to evolve alongside Digital Sportsman’s data model as the business grows. The MCP-based tool architecture supports modular expansion, as new API capabilities can be added by registering additional tools without modifying the core agent logic. Avahi’s rapid iteration capability was demonstrated early in the engagement when a critical chatbot tone issue was identified and resolved within 24 hours through prompt tuning, establishing a pattern of responsive delivery that continued throughout the project.
Let’s explore your high-impact AI opportunities together in a complimentary session
