The Shift from Legacy Dashboards to Conversational AI Analytics
Building AI-powered analytics products at enterprise scale isn't just challenging—it's a race against rapidly evolving customer expectations. At the recent Boundaryless Product Launch event, we interviewed Aria Moshari, Director of Software Engineering at Verisk.
He shared how his team transformed their approach to delivering AI analytics products, cutting development time from months to just 12 weeks while managing 30 petabytes of data across insurance and financial services.
The 80/20 Reality: Why Data Quality Drives Trustworthy AI
For business and data leaders wrestling with AI transformation priorities, Verisk's experience offers a crucial insight about where to focus your efforts first.
"20% of our work involves AI technologies, but the other 80% of the work all relies on the data itself– the trustworthiness, depth, and the quality of the data," Moshari explained during the interview.
This perspective shifts the conversation away from chasing the latest AI models toward building the data foundation that makes AI successful. For enterprises managing massive data volumes like Verisk's 30 petabytes, this foundation becomes even more critical.
Key insights from Verisk's data-first approach:
Data quality drives trustworthy AI: While teams often focus on algorithms and models, 80% of successful AI implementation depends on trusted, high-quality data.
Scale demands strategic partnerships: At enterprise levels, internal teams need external expertise to handle the complexity effectively.
Foundation first, AI second: Organizations excel when they master data management before layering on AI capabilities.
"We are reimagining Verisk to be an AI-first organization at scale," Moshari noted, emphasizing that their strategy centers on partnerships with best-in-class providers rather than building everything internally.
💡See how leading teams are delivering branded, interactive intelligence directly inside their products—watch our biggest product launch yet here.
Build or Buy? How the Right Partnership Delivers Faster Time-to-Market
Product leaders face a critical decision that directly impacts time-to-market: build or buy? Verisk's experience provides a clear framework for making this choice.
The changes happening now differ from previous technology shifts in one crucial way—speed.
As Moshari observed, "This transformation is happening at a much faster pace; we didn't have that experience before or maybe very rarely."
Customer expectations have evolved beyond traditional search and filtering. "Right now, their expectation is just to type their question in their mind and get the answer instantly," Moshari explained.
Strategic considerations for product leaders:
Time-to-market pressure: The question isn't whether to build AI-powered products, but when you can deliver them to customers.
Competitive benchmark movement: One product leader mentioned building in-house for five years, but never succeeded due to advancing competitive standards.
Resource allocation reality: Organizations operating at scale need help to meet rapidly evolving user expectations.
Partnership necessity: As Aria says, "That journey, you cannot take it alone. You need a stronger strategic partnership."
The choice between 2-4 months versus 12 months to market can determine whether you lead or follow in your industry.
Embedded Analytics at Enterprise Scale with Spotter
When Verisk evaluated analytics providers, they focused on specific capabilities that would accelerate their AI integration while handling complex enterprise requirements.
Their evaluation criteria centered on embedded AI capabilities, conversational BI, powerful data insights, complex data model handling, and seamless integration processes. ThoughtSpot’s capability for embedding AI agents like Spotter proved especially critical for development efficiency.
"Our experience was using Spotter and then embedding it in our product; we were not talking about months of work. We were talking about weeks," Moshari shared.
Concrete results from Verisk’s embedded AI implementation:
12-week product delivery: Verisk went from ground zero to production in just twelve weeks for a complete AI-powered analytics product.
Seamless embedding: The integration process significantly reduced development costs and effort.
Complex data model support: The platform handled Verisk's sophisticated enterprise data requirements.
Conversational BI capability: Users can ask questions naturally instead of navigating complex interfaces.
The platform's ability to handle both the technical complexity and user experience requirements made it possible for Verisk to achieve its aggressive timeline goals while maintaining enterprise-grade functionality.
💡 Want your own secret weapon for smarter SaaS products? Get your copy of the Embedded AI Agents ebook here.
Evolving Embedded Analytics: Predictive, Conversational, and Real-Time
The shift in user expectations represents more than a UI change—it's a fundamental transformation in how people interact with data and what they expect from analytics platforms.
Moshari highlighted this evolution: "I'm very excited to see the transformation from reporting to the point that we can start predicting, actually, and from the legacy dashboard to the real conversations."
This boundaryless transformation addresses key pain points for business leaders who need faster, more intuitive access to insights, and data leaders who want to reduce the bottleneck of report requests.
Future capabilities for Verisk:
Unstructured data expansion: Moving beyond structured data to include unstructured data sources
Predictive capabilities: Evolving from historical reporting to predictive analytics
Natural conversations: Replacing traditional dashboard navigation with conversational interfaces
Real-time insights: Instant answers to business questions without waiting for report generation
"With partners like ThoughtSpot, we can do this both at scale and at speed," Moshari concluded, emphasizing how the right partnership enables both the technical capabilities and the rapid delivery timeline that modern businesses require.
How to Win with an Embedded Analytics Strategy Like Verisk
The question isn't whether to transform your analytics approach, but how quickly you can deliver AI-powered insights that keep pace with your market.
Verisk provides a clear roadmap: Focus on data quality first, evaluate partnerships based on speed and embedding capabilities, and prioritize conversational user experiences that meet evolving customer expectations.
Book your demo today to explore how embedded analytics can accelerate your time-to-market.
