AI Agents in Snowflake for Data Analysis
Drive Growth and Retention by Converting Customer Interactions into Actionable Priorities
Written by Enrique Fueyo, Jorge Penalva, founders of Lang.ai, and Cameron Wasilewsky, Senior Sales Engineer at Snowflake
Snowflake has revolutionized data consolidation, but transforming this wealth of information into business outcomes remains a challenge. Lang.ai addresses this with AI agents that analyze customer data in Snowflake. Our solution deploys custom AI models that connect unstructured feedback with key variables like customer type and revenue. These agents transform customer interactions into actionable priorities, delivering insights that drive retention and growth strategies.
Who We Are
We are Enrique and Jorge, founders of Lang.ai, our second startup together. Our first company, a social media analytics platform for the Spanish-speaking market, grew from Spain to Latin America, serving leading enterprise customers in the region. In 2018, we decided to start our second company, Lang.ai, in San Francisco to address the exponential growth of digital customer interactions. The digital transformation of businesses brought customer communications and actions online, generating vast amounts of data with untapped potential for insights.
Now, as enterprise data increasingly consolidates in the Snowflake AI Data Cloud, we’re able to fully realize our vision. Building in this ecosystem allows us to provide deeply contextualized analysis that directly ties to business outcomes.
The Problem We’re Solving
In today’s digital landscape, companies struggle to extract meaningful insights from large amounts of customer interaction data. Traditional Voice of the Customer (VoC) platforms only scratch the surface, consolidating feedback without connecting it to its impact to the revenue of the company. This leaves product and other business teams unable to make truly informed decisions based on customer data.
By having access to enterprise data in Snowflake, Lang.ai can overcome this limitation, using LLMs to connect customer feedback with key company metrics and deliver actionable insights directly to decision makers.
Our Solution: AI Agents for Analytics of Customer Interactions
Lang.ai’s Snowflake AI Agents extract actionable priorities from customer interactions, directly impacting business metrics like retention, onboarding conversion rates and revenue long term. Our AI agents:
- Automate complex and time consuming data engineering for unstructured data.
- Bring normally separate data from product, customer interactions and revenue together into one clear story.
Our solution seamlessly integrates with your Snowflake data, enabling data teams to build and deploy these AI-powered analytics agents for business teams without the need for complex setup or AI expertise.
- Snowflake data processing: Our AI agents analyze customer data within Snowflake, including feedback, purchase history, digital event data and buying behavior, uncovering insights that might otherwise go unnoticed.
- Contextual analysis: Our agents go beyond surface-level analysis by connecting unstructured feedback with structured business data. They can link customer complaints about a product to specific user segments and their buying behavior. For example, they may identify a recent increase in churn rates of premium customers and note that 20% of them had a bad experience with product XYZ’s packaging.
- Continuous proactive prioritization: The AI agents constantly analyze incoming data, extracting the most relevant priorities to focus on and revealing hidden correlations and trends. In the previous example, this means the company can now focus specifically on packaging as a churn factor and decide whether they should temporarily discontinue product XYZ or keep it on the shelves while fixing its packaging issues.
- Integration with existing business workflows: Our AI agents integrate with your existing workflow tools, such as Slack, ensuring that insights are delivered in real-time to the right decision-makers, facilitating quick and informed action.
So, How Do You Set an AI Agent to Improve Retention for Retail/ CPG?
Here’s how our AI Agent can drive impact for a retail CPG company looking to improve customer retention:
- Connect feedback to outcomes: The AI agent analyzes customer feedback alongside purchase history and product data, linking issues like quality concerns or delivery delays to specific product lines or operational processes.
- Prioritize actions: Based on these insights, the agent identifies high-impact areas for improvement, such as optimizing the supply chain for frequently delayed products or enhancing quality control for items with consistent complaints.
- Measure and adapt: The agent continuously monitors customer behavior, purchase patterns and feedback, tracking the impact of implemented changes and dynamically adjusting recommendations to maximize retention.
The Setup Process, Step by Step
Step 1. Install Lang.ai app from Snowflake Marketplace
Search for Lang.ai on Snowflake Marketplace; from the app listing page, click on “Get” to access.
Step 2. Create a view with your data
Prepare your data for analysis by creating a custom view in Snowflake. Below is an example of a view for artificially generated dataset of Nike which contains about 1000 survey responses for lapsed customers that stopped buying after being frequent monthly buyers.
Step 3. Set up the app with your view
Configure the Lang.ai app to use your specific data view for analysis.
Step 4. Launch the Retention AI agent
Activate the AI agent to begin analyzing your customer data and generating insights.
Step 5. Navigate your agent insights in Slack and in the Snowflake Native App’s UI
Access, interact and filter the AI-generated insights via Slack and the Snowflake Native App’s UI.
“Implementing this AI agent for Retail/CPG can significantly boost retention by identifying at-risk customers early, personalizing engagement strategies and optimizing product offerings based on customer behavior and feedback. This leads to reduced churn, increased customer lifetime value and more efficient allocation of resources.”
Technical Deep Dive
Our application architecture leverages the Snowflake Native App Framework to deliver AI-driven insights directly in Slack. The architecture includes:
- Agent Configuration: Users can select and configure AI agents based on specific goals.
- Agent Orchestration: This component pulls data from the Snowflake data warehouse for analysis.
- Output to Slack: Insights are delivered directly to product managers & other business teams via Slack.
- Agents Web UI: The web UI provides detailed context and graphical support.
The Snowflake Native App Framework has been instrumental in enabling our AI capabilities by providing:
- Comprehensive analysis: Consolidates diverse data sources, allowing our AI agents to access all relevant information for contextual insights.
- Direct linking of insights to business outcomes: Connects customer interactions directly to business outcomes, driving actionable recommendations with real impact.
- Seamless deployment and integration: Enables secure AI model deployment and easy integration with existing tools, ensuring high adoption rates.
Valuable Lessons and Future Vision
Our journey with Lang.ai required a bold pivot as the AI market rapidly evolved. Transitioning from a multi-tenant SaaS to Snowflake’s single-tenant environment was challenging, but it unlocked unprecedented potential for secure, powerful AI analysis. This shift allowed us to overcome the limitations and risks associated with consolidated customer data in traditional SaaS models.
Snowflake’s robust infrastructure, particularly Snowpark Container Services and user-defined functions, has been instrumental in enhancing our AI agents’ capabilities. We’ve leveraged these tools to deliver advanced LLM-powered data analysis with unparalleled security and insights depth.
Looking ahead, we see AI as a catalyst for agile, high-impact business operations. This vision, coupled with our commitment to deliver value to customers, allows us to make AI truly useful.
Lang.ai’s AI Agents for Snowflake represent a significant step forward in how businesses can leverage their data. By utilizing the Snowflake Native App Framework, we’ve created an LLM-based solution that:
- Integrates seamlessly with existing data infrastructure
- Requires minimal setup and maintenance
- Operates directly on your Snowflake data, supporting data governance and security
- Provides deep, contextual insights from both structured and unstructured data sources
Click here to access Lang.ai on the Snowflake Marketplace and start your free trial today!
