Vibe Querying Episode 8: Vibing for Product Management Intelligence

Product managers use CData's Salesforce MCP Server to transform unstructured sales data into structured use cases, enabling data-driven product roadmap decisions and automated Salesforce record creation.

The eighth episode of our Vibe Querying series takes us back into the world of product management, where we explore how product managers can leverage CData's Salesforce MCP Server to address sales data hygiene and extract customer use cases for product roadmap planning. Joining us for this episode is Jonathan Hikita, Director of Product Management at CData Software, who demonstrates how he uses MCP to transform weeks of manual analysis into hours of conversational data exploration.

Watch Now: Vibe Querying with MCP - Episode #8

Introducing MCP, CData MCP Servers, and Vibe Querying

Model Context Protocol (MCP) is a protocol developed by Anthropic that defines how AI agents and LLMs can communicate with external systems securely and efficiently. MCP enables AI models to access real-time data and interact naturally with that information through conversational AI or agent-driven actions.

CData MCP Servers connect CData's connector library of 350+ data sources to your favorite AI clients like Claude or Gemini. This combination gives AI models access to the work data they need to be genuinely useful in business contexts. In this episode, we're showcasing both local MCP servers and remote MCP capabilities through our Connect AI product.

This brings us to the concept of "vibe querying" - much like vibe coding, it's a conversational approach to data exploration. You can get information about your opportunities, close tickets, build dashboards, and prescribe actions based on business system insights—all without writing a single line of code.

Today's focus: Product management intelligence and data hygiene

For today's episode, we're tackling two critical challenges that product managers face: addressing sales data hygiene by ensuring documented use cases for all opportunities and understanding customers' business challenges to inform product roadmap and marketing decisions.

Jonathan manages CData's drivers-level products, which present a complex challenge: supporting 270+ data sources as drivers, close to 200 read-only profiles, and over 200 tested data consumers. This creates approximately 100,000 possible combinations across thousands of customers and OEM implementations. The key question: which scenarios should the product team focus on for maximum customer impact?

Traditional approaches require manually reviewing Salesforce opportunities, reading through email conversations, analyzing meeting notes, and compiling use case information—a process that typically takes weeks to cover just one month of data.

The setup: Custom Salesforce objects and unstructured data

Before diving into queries, Jonathan explained the data architecture: "We use a custom use case object in Salesforce that's typically associated with an opportunity. But in many cases, use case information is scattered in email conversations, sales rep notes, and support case information. We need to extract data from those unstructured sources."

This represents a perfect example of how AI can bridge the gap between structured systems (Salesforce custom objects) and unstructured business intelligence (notes, emails, activities) that would be impossible to capture efficiently through traditional reporting.

The queries and insights

Note: Any numbers and customer names presented below have been obfuscated (by a very neat Claude instruction).

Query 1: Identifying opportunities without documented use cases

The question

"From the latest Salesforce opportunities, find 10 opportunities that are new business and have no use case object associated."

What Claude did

• Explored Salesforce data structure to understand opportunity and custom use case object relationships

• Filtered opportunities by business type (new business) and missing use case associations

• Generated a prioritized list of opportunities requiring use case documentation

• Identified specific opportunity names like "JDBC Salesforce," "QuickBooks connecting from Power BI," and "SAP HANA to Databricks"

The impact

Claude immediately identified the data hygiene gap that Jonathan needed to address. However, as Jonathan noted, "These opportunity names don't really explain the full use case." This set up the perfect transition to extract the missing intelligence from unstructured data sources.

Query 2: Extracting structured use cases from unstructured data

The question

"Read email messages from activities, opportunity descriptions, and cases to identify the use cases of these 10 opportunities. Use cases should have data source information, connecting technology, data consumer/destination, and the purpose."

This represents what Jonathan calls "purposeful prompting"—being very specific about what information to extract and how to structure it.

What Claude did

• Analyzed email conversations associated with each opportunity

• Parsed meeting notes and activity descriptions

• Extracted sales representative notes and comments

• Cross-referenced support case information when available

• Synthesized unstructured text into structured use case components

The impact

Claude transformed scattered, unstructured information into comprehensive use case intelligence:

Sample extracted use cases

• CData Sync (ETL solution): SAP HANA → Databricks for analytical purposes

• HubSpot → BI tools for reporting and analytics

• QuickBooks → Power BI for analytical use cases

• QuickBooks Online → Power BI for business intelligence

As Jonathan reflected on the efficiency gain: "What I used to do before I could use Claude and MCP was create a report that doesn't have use case objects, then go look at every conversation and activity that happened with the customer to capture this. It would take me weeks to cover what happened in one month. With Claude, I only spent a few hours."

Query 3: Writing structured data back to Salesforce

The question

"Write back the number five use case information to the use case object associated with Salesforce."

What Claude did

• Retrieved the custom use case object schema and field definitions

• Analyzed existing use case records to understand dropdown values and data formats

• Validated field requirements and acceptable input values

• Created a properly formatted use case record in Salesforce

• Provided confirmation with the new record ID

The impact

This demonstrated the bi-directional capabilities of CData MCP Servers. Jonathan became the first guest in the Vibe Querying series to successfully write data back to a source system, transforming the extracted intelligence into actionable business data that the entire team could leverage.

Key technical achievement

• Automated data hygiene improvement

• Institutional knowledge preservation

• Team-wide access to structured use case intelligence

• Elimination of manual data entry workflows

Query 4: Strategic product insights and market trends

The question

"As a product manager, what should I get from these 10 use case information?"

What Claude did

Without requiring additional data queries, Claude synthesized the use case intelligence into strategic product insights:

• Identified market trend patterns across the 10 opportunities

• Calculated percentage breakdowns of popular source-destination pairings

• Highlighted industry timing factors affecting customer needs

• Connected use case trends to broader market dynamics

The impact

Key strategic insights revealed:

Market timing opportunity:

• 30% of use cases involved QuickBooks → Power BI integration

• Microsoft Power BI is deprecating the native QuickBooks connector

• CData is a trusted Microsoft partner positioned to support affected customers

Popular integration patterns:

• Two HubSpot customer opportunities indicating CRM analytics demand

• SAP HANA → Azure Databricks representing enterprise data warehousing trends

• Multiple Power BI destinations showing Microsoft ecosystem adoption

Actionable business intelligence:

As Jonathan noted: "The first insight is exactly what I needed to know as a product manager. There are users in need because of industry changes. We have the capability to support them, and we need to take action in marketing and sales to help customers find our solution and make the transition as smooth as possible."

The business impact: From weeks to hours

The transformation in product management workflow was dramatic. Traditional use case analysis required:

• Creating Salesforce reports for opportunities without use case objects

• Manually reviewing every customer conversation and activity

• Reading through email threads and meeting notes

• Compiling findings into structured intelligence

• Weeks of effort to analyze one month of opportunities

With the MCP-powered approach, Jonathan can now:

• Generate comprehensive use case intelligence in hours instead of weeks

• Automatically extract insights from unstructured data sources

• Write structured findings back to Salesforce for team access

• Focus on strategic decision-making rather than data compilation

• Identify market timing opportunities immediately

Technical excellence: Universal database interface with custom objects

One of the most impressive technical aspects was how CData MCP Servers handled Salesforce custom objects with the same sophistication as native objects. The system demonstrated:

Deep Salesforce integration

• Access to all custom objects and fields (identifiable by "_c" suffixes)

• Understanding of custom dropdown values and field validation rules

• Bi-directional data operations (read and write) with custom objects

• Complex relationship navigation between opportunities and custom use case objects

Intelligent data processing

• Self-correction when encountering API errors or formatting issues

• Contextual understanding of business relationships within Salesforce

• Validation of data formats before writing back to the system

• Preservation of data integrity across read-write operations

As Marie highlighted: "This is a unique capability of CData connectivity. Because we have such deep and robust connectivity with Salesforce, we can handle data much better with complex database functions and system calls across large datasets for both read and write operations."

What this means for product management teams

This episode showcases how AI-powered use case analysis can transform product management:

Speed: Compress weeks of manual analysis into hours of conversational exploration

Coverage: Analyze unstructured data sources that would be impossible to review manually at scale

Actionability: Transform scattered intelligence into structured, shareable business insights

Market responsiveness: Identify timing opportunities and industry trends immediately

Team collaboration: Create structured data accessible across sales, marketing, and product teams

From scattered intelligence to strategic insights in hours

As Jonathan summarized the transformation: "This allows the entire team to benefit from use case information. As a product manager, I can now focus on what I should do with the insights rather than spending weeks gathering them."

The ability to seamlessly extract use case intelligence from unstructured sources while simultaneously populating structured systems represents a fundamental shift in how product teams can operate. Instead of choosing between speed and thoroughness, MCP enables both.

Getting started with your own use case analysis

If you want to unlock similar insights for your Salesforce product intelligence:

1. Download the free beta: Visit cdata.com/solutions/mcp to access the Salesforce MCP Server

2. Remote MCP capabilities: Try Connect AI at cdata.com/cloud for multi-source access

3. Simple installation: Follow the automated installer for Claude Desktop integration

4. OAuth connection: Authenticate with your existing Salesforce credentials

5. Start conversations: Explore the CData prompt library for proven queries and begin asking strategic questions about your customer data

For more examples of vibe querying in action, explore our complete episode series covering sales, marketing, project management, and more use cases.

Beyond manual data compilation: The shift to conversational product intelligence

This episode demonstrates that we're moving beyond manual use case compilation and into conversational product intelligence. When AI can access, analyze, and act on your customer data through natural language, product managers become strategic orchestrators rather than data archaeologists.

Product teams can now ask the complex questions they've always wanted to explore: Which use cases are emerging fastest in our pipeline? What industry timing factors are creating new opportunities? How can we better predict and respond to market changes?

For deeper insights into how MCP is transforming product management and customer intelligence, explore our comprehensive blog series on MCP implementation strategies and use cases.

Join us for future episodes of Vibe Querying with MCP, where we continue exploring how natural language AI transforms business intelligence. Until next time, stay curious and keep vibing.