Transforming Customer Experience with Snowflake: The Tasty Bytes Story
Harnessing the AI Data Cloud to Revolutionize Customer Support and Operational Efficiency
Customer Experience is at the core of any business. Tasty Bytes is a fictitious global food truck brand that leverages Snowflake to enhance many facets of customer experience. In this blog we will showcase how Tasty Bytes uses Snowflake to improve Customer Experiences through Unstructured Data Extraction, Review Translation and Summarization, and RAG based Chat and Email Support. We will accomplish all of this while leveraging cutting edge Snowflake functionality including Cortex large language models (LLMs), Streamlit in Snowflake, Document AI, Snowflake Notebooks and more.
Tasty Bytes Overview
Tasty Bytes is one of the world’s largest food truck networks, operating in 30 major cities across 15 countries with 15 core brands. They offer unique, high-quality food options that are safe, convenient, and cost-effective. By sourcing ingredients from mostly local vendors, they ensure their success positively impacts community partners.
While Tasty Bytes has ambitious goals, they had issues with Customer Experience. In particular Tasty Bytes had the following challenges:
- Cumbersome Customer Support: Understanding customers and food trucks is a manual process. This includes reading reviews, chat logs, and inspection PDFs from multiple sources. Support agents typically lack a full picture of customers, which leads to customer frustration.
- High Volume of Customer Tickets: The email support queue is flooded with basic questions. This causes longer response times and customers frustration. The high volume of basic questions reduces agents’ ability to accurately respond to complex questions.
- High Support Rep (Live Agent) Turnover: Agents do not have the tools and information they need to succeed. This is leading to high turnover and continuously onboarding new agents with lengthy ramp up, negatively impacting service consistency and quality.
- Fragmented Customer Insights: Difficulty in consolidating feedback from diverse channels such as social media, online reviews (Yelp, TripAdvisor, etc.), and direct interactions. Additionally, reviews span a multitude of languages, making it difficult to have a holistic view of customer feedback.
Let’s dive deeper into how Tasty Bytes solved these issues using Snowflake.
AI Driven Insights and Automation
Equipping Tasty Bytes support agents with comprehensive information access enables them to handle customer inquiries more effectively. The Snowflake-powered platform reduces mundane tasks, decreases ramp time, and significantly enhances their effectiveness and job satisfaction.
Tasty Bytes developed a Streamlit in Snowflake application to streamline call center workloads by automating email responses using a knowledge base. Streamlit is an open-source Python library that makes it easy to create and share custom web apps for machine learning and data science. Streamlit in Snowflake helps developers securely build, deploy, and share Streamlit apps on Snowflake’s data cloud.
When an email arrives into the application, a Snowflake Cortex LLM assesses if the knowledge base can address the query. Snowflake Cortex gives you instant access to industry-leading LLMs that are fully hosted and managed by Snowflake and using them requires no setup. Your data stays within Snowflake, giving you the performance, scalability, and governance you expect.
If the Cortex LLM determines it can answer the email from the knowledge base, an automatic response is sent; if not, the email is forwarded to an agent. Agents can choose to reply to customer emails, with the Cortex LLM providing suggested responses. They can also contribute responses to the knowledge base, allowing the Cortex LLM to handle similar queries automatically in the future.
For a more detailed overview and implementation instructions, check out the Tasty Bytes — Customer Support Streamlit Application Powered by Cortex Quickstart.
Understanding Our Customer with Customer Review Analytics
With Snowflake as our single source of truth, Tasty Bytes ran cross data set analysis to pinpoint where our customer experience may be lacking at the truck and operations levels. This allowed Tasty Bytes to personalize interactions, predict customer needs, and inform food truck owners about improvement areas.
Tasty Bytes used Snowflake Cortex within Snowflake Notebooks to identify shortcomings in their customer experience at both the truck and business levels. Snowflake Cortex provides LLM-based functions such as summarization, sentiment analysis, and complete (prompt responses) to understand unstructured data, answer freeform questions, and provide intelligent assistance.
Tasty Bytes leverages Snowflake Cortex to analyze customer reviews from multiple sources and languages to get feedback on their food trucks. Their comprehensive analysis in Snowflake helps them understand areas needing improvement, enhancing customer loyalty and satisfaction.
For a more detailed overview and implementation instructions, check out the Customer Reviews Analytics using Snowflake Cortex Quickstart.
Empower Agents with RAG based Chatbot
Tasty Bytes implemented a Snowflake Cortex LLM powered chatbot to help agents find answers to customer questions faster. This Streamlit in Snowflake chatbot uses retrieval augmented generation (RAG) by retrieving previous agent support chat logs and public-facing documents (Annual Reports, FAQs) stored in Snowflake tables.
Customers are happier because agents have the information needed to provide accurate, timely responses, and LLM-powered chatbots ensure quick and consistent support. This leads to faster solutions, reduced wait times, more personalized service experiences, and higher rates of first contact resolution.
For a more detailed overview and implementation instructions, check out the Tasty Bytes — RAG Chatbot Using Cortex and Streamlit Quickstart.
Operational Excellence with AI Powered Truck Analytics
Tasty Bytes developed a Snowflake Cortex powered Streamlit in Snowflake application to improve customer experience by analyzing customer reviews and food truck inspections. Within one application, they enabled management with truck and customer review data, including inspection PDFs.
It streamlines communication with truck owners through sentiment analysis and assists in drafting emails. The app includes built-in analytics for users to interact with review data and generate plots using LLM capabilities. Document AI integration further enhances analysis by extracting insights from handwritten, unstructured inspection documents.
For a more detailed overview and implementation instructions, check out the Tasty Bytes — Enhance Customer Experience Streamlit App Quickstart.
Drive Insights from Unstructured Data
Document AI gives Tasty Bytes the ability to drive insights from unstructured data. Document AI is a Snowflake AI feature that uses a proprietary LLM to extract data from documents. Food trucks generate inspection reports that they convert to structured data, ready to analyze. This allows agents and analysts to have a full picture of trucks, including data that was previously underutilized.
For a more detailed overview and implementation instructions, check out the Extracting Insights from Unstructured Data with Document AI Quickstart.
Unlock all Enterprise Data
Tasty Bytes has customer reviews living outside Snowflake in Iceberg tables. Using Iceberg in Snowflake, they discovered insights by combining data inside and out of Snowflake. This enables them to fully understand customers and ensure they are meeting their needs. Tasty Bytes even used Cortex for sentiment analysis using review data in Iceberg.
For a more detailed overview and implementation instructions, check out the Tasty Bytes — Working with Iceberg Tables Quickstart.
Conclusion
In conclusion, Tasty Bytes has successfully leveraged Snowflake’s advanced functionalities to revolutionize their customer experience. By addressing challenges in customer support, streamlining communication, and harnessing AI-driven insights, they have enhanced customer satisfaction and operational efficiency. You can achieve similar success by adopting these strategies and tools. Explore how Snowflake’s capabilities — such as Cortex LLMs, Streamlit in Snowflake, and Document AI — can transform your customer experience. Start implementing these solutions today to drive customer loyalty and operational excellence.
