The Trucker Media Group (TTMG) sought to modernize its digital driver recruitment funnel, moving away from impersonal, static forms to build a cloud-native solution that actively reduces applicant friction and improves job application conversion rates. SnapSoft delivered a production-grade, fully serverless architecture on AWS, engineered around a sophisticated conversational AI agent. This solution features a stateful, agentic workflow that intelligently orchestrates the entire user journey from discovery to application.

This agent, powered by Amazon Bedrock, goes beyond simple Q&A. It autonomously determines user intent, manages complex conversational context over long-running sessions, and proactively performs actions on the user's behalf. Key agentic behaviors include dynamically parameterizing and querying TTMG's external Job API to conduct real-time job searches, contextually answering specific questions about a selected job, and intelligently guiding the user through the "initial application." Once the agent determines all required information has been collected, it automatically orchestrates the complete lead submission process to TTMG's backend. The result is a secure, resilient, and highly scalable workflow that transforms the driver experience, significantly reducing applicant drop-off and maximizing lead conversion rates.

The Trucker Media Group (TTMG) needed to reduce significant friction in its online driver recruitment funnel and improve conversion rates between its initial "short form" and the final, DOT-compliant "long form" application. The existing static, form-based process struggled with:

• High Applicant Drop-off: A lack of real-time, personalized engagement caused interested drivers to abandon the application process.
• Static User Experience: The platform could not contextually answer a driver's specific questions about a job posting they were viewing, leading to a disconnected and impersonal journey.
• Inefficient Lead Capture: The system was unable to conversationally guide users or capture partial information, resulting in lost opportunities from qualified leads who were interested but not ready to complete a full job application.
• Lack of Scalable Engagement: Providing 24/7, real-time, one-on-one support to thousands of drivers simultaneously was not operationally feasible with the existing setup.

TTMG required a modern, scalable solution that could provide an engaging, AI-powered conversational experience to intelligently guide drivers, answer their questions 24/7, and seamlessly capture their information to accelerate applicant conversions.

SnapSoft designed and implemented a fully serverless, conversational AI platform to serve as a "sidecar" component, providing an intelligent, scalable, and engaging experience for TTMG's driver recruitment process.

• Generative AI Conversational Core: The solution uses Amazon Bedrock as its intelligent core. It employs a multi-model strategy, using Nova Lite for rapid intent detection and Anthropic Claude 4.5 to handle the complex, "agentic-like" conversational workflow, guide users through the short form, and answer context-specific questions about job postings.
• Real-time Job Search & Lead Submission: The central AWS Lambda (Prompt Orchestrator) acts as the system's brain. It securely accesses AWS SSM Parameter Store for credentials to query TTMG's external Job API in real-time. Once the user is qualified, the Lambda automatically submits the lead (short form data) to TTMG's backend systems.
• Stateful Context Management: Amazon DynamoDB provides a high-performance, persistent session store. This allows the chatbot to maintain full conversational context, store collected user attributes, and recognize returning users, ensuring a seamless and personalized journey.
• Secure and Resilient Architecture: The application is deployed in a secure, multi-AZ VPC. The Private REST API Gateway and an API GW Interface Endpoint ensure TTMG's webserver communicates with the chatbot without traversing the public internet. The core Lambda runs in private subnets, accessing Bedrock via a VPC endpoint and the external Job API via a NAT Gateway.
• Infrastructure as Code: The entire solution, from the networking to the serverless functions, is defined and provisioned via Terraform. This IaC approach is managed in a Bitbucket pipeline, enabling automated, version-controlled, and reproducible deployments.

AWS was chosen for its leadership in managed Generative AI, its robust serverless-first ecosystem, and its comprehensive security and integration capabilities.

Its managed GenAI service, Amazon Bedrock, along with serverless components like AWS Lambda, Amazon API Gateway, and Amazon DynamoDB, provided the complete foundation for building and automating TTMG's intelligent, stateful driver recruitment workflow. With multi-AZ deployments for resilience, private networking (VPC Endpoints and Private API Gateway) to securely connect to TTMG's existing application, and IaC-based provisioning, AWS offered the scalability, security, and enterprise-readiness required for a public-facing solution.

• Amazon API Gateway (Private REST endpoint for production integration, Public REST endpoint for demo/debugging)
• AWS Lambda (Serverless "Prompt Orchestrator" function containing the core agentic workflow and business logic)
• Amazon Bedrock (Managed GenAI service providing on-demand access to Anthropic Claude and Nova Lite models)
• Amazon DynamoDB (High-performance, serverless database for session store, chat history, and user attributes)
• Amazon VPC (Isolated network with private/public subnets, NAT Gateway, Interface Endpoints, and Gateway Endpoints)
• AWS WAF (Web Application Firewall used for IP whitelisting to secure the public demo API Gateway)
• AWS SSM Parameter Store (Secure storage and management for credentials, such as external API keys)
• Amazon S3 (Object storage for logs and enabling downloadable user chat transcripts)
• Amazon CloudWatch (Monitoring, logging, and observability for all serverless components and application performance)
• AWS IAM (Identity and Access Management for defining secure, least-privilege roles and policies for all AWS services)
• Terraform (Infrastructure as Code (IaC) tool used for automated, reproducible provisioning of all AWS resources)
• Bitbucket (CI/CD platform used to host the Terraform code and run deployment pipelines)