Solution
Hershey Entertainment & Resorts built a comprehensive customer and employee engagement platform using BigQuery, Gemini, Conversational Agents, and the Customer Engagement Suite IVR ('Agnes'). Guest service data from calls and web activity aggregates into BigQuery, where Gemini analyzes sentiment, engagement surveys, and interaction patterns. Conversational Agents power a chatbot deployed on the website within 40 hours to handle invoice requests and password resets. The Customer Engagement Suite deploys 'Agnes', an agentic IVR system that recognizes callers, automates employee call-offs (25%+ of volume), and manages password reset requests (50%+ of calls) via webhook integrations to internal systems.
Data flow
Guest calls and web interactions flow into Google Analytics and BigQuery. Call metadata and IVR interactions captured by Customer Engagement Suite feed into BigQuery. Gemini analyzes aggregated data to identify patterns and sentiment. Website visitor behavior triggers Conversational Agents for chatbot responses. Employee or guest calls route through 'Agnes' which queries internal systems via webhooks to determine action (call-off, password reset, invoice request) and responds conversationally.
Solution architecture
7 components · 4 layers - Compute
- Gemini Analyzes complete call and engagement datasets to identify sentiment trends, summarizes employee engagement surveys in seconds, and provides actionable recommendations for guest experience improvements.
- Dialogflow Enables natural language conversation flow design for both website chatbot and IVR interactions.
- Storage
- BigQuery Centralizes guest call data, web activity from Google Analytics, social sentiment from Facebook forums, and employee survey responses for comprehensive analysis.
- Serving
- Conversational Agents Powers the website chatbot that handles guest invoice requests and password resets with dynamic, segment-aware responses.
- Customer Engagement Suite Deploys 'Agnes' IVR agent that recognizes callers, orchestrates call-offs, manages password resets, and identifies customer needs via webhook integrations to backend systems.
- Looker Studio Generates real-time reports from Dialogflow and BigQuery data, accessible to managers for call volume and performance monitoring.
- Uncategorized
- Google Analytics Tracks guest web sessions and engagement to feed into BigQuery for comprehensive customer journey analysis.
- 'Agnes' IVR on Customer Engagement Suite uses caller-ID webhooks to identify callers and read context (account lockouts, failed logins, open support tickets) from HE&R's enterprise systems
- Agnes proactively suggests password reset emails to locked-out users rather than presenting an IVR menu
- Dialogflow interactions tagged with three answer fields are pushed to BigQuery and visualized in Looker Studio
- Unified customer interaction data stored in BigQuery as the single analytical store
- Vertex AI and Gemini summarize social-channel sentiment (e.g. the 'Connoisseurs' Facebook page) and the annual employee engagement survey
- Website chatbot powered by Conversational Agents extended with a webhook that takes an email address, checks for stays in the last 60 days, and emails a PDF invoice
Fully automated more than 25% of team member 'call off' requests
Implemented a natural language customer chatbot in less than 40 hours
Manages 50%+ of password reset calls with Customer Engagement Suite
Since Agnes is plugged into our infrastructure, it can see things like account lockouts, failed logins, or open support tickets
That's a full eight hours of employee work per day. By saving those hours, our people can spend time innovating and solving more complex guest problems.
We can take our Dialogflow interactions, push them to BigQuery by answering three questions in Dialogflow, and then go to Looker Studio to create a report in 30 minutes.
