Resource hub

Test Driving GPT-5.2

OpenAI released GPT-5.2 in December 2025, promising improvements in reasoning, instruction following, and complex query handling over previous versions such as GPT-5.1 and GPT-4.1. We test drove GPT-5.2 for Tursio’s structured data search and summarize our findings in this blog. 

Structured Data Search with Microsoft Context

Microsoft is a trusted provider across various industries, and given our core members spent years building Microsoft products, it is natural for us to treat Microsoft as a first-class citizen in our structured data search.

The Tursio Journey

Data is to businesses like corals are to seas -- critical, fascinating, and full of potential. Just like corals, data systems and technologies have stayed relevant even as the world kept looping through successive technology cycles, including mainframe, desktop, web, mobile, cloud, and now AI. Shi and I saw this opportunity with data transitioning into AI when we started Tursio.

From data modeling to knowledge graphs

Data modeling is essentially the process of building rich business context on top of raw data. However, building this context manually requires significant effort. A knowledge graph can subsume data modeling by providing the relevant context dynamically.

Navigating LLMs: Is newer the better for data tasks?

OpenAI released GPT-5 earlier this year and the tech community immediately jumped to upgrade their AI systems. This was a natural reaction given that every new GPT model has gotten substantially better over the years.

SQL Server 2014: A Milestone That Still Powers the World’s Data

While the rest of the world is racing toward the latest AI-native databases and cloud data platforms, I find myself reflecting on a product that’s now over a decade old — SQL Server 2014.

Why Search Scales Differently from Data or BI

Search box powered by AI is quickly becoming the new expectation for accessing enterprise data. However, interestingly, this search box can reside in multiple places: it can be right within the database platform where the data is stored, or in the BI tool where the dashboards are built, or be standalone like Tursio. In this blog, we discuss the scalability aspect with these different architectural choices.

From 80% to 100%: The Accuracy Marathon in Enterprise AI

When organizations evaluate AI solutions, accuracy often becomes the headline metric. But what those numbers don’t reveal is the story behind them: the people, processes, and persistence required to transform an engine from “good enough” to exceptional. This is critical since most AI tools struggle with accuracy and are anywhere between 50-80% accurate out-of-the-box, which is typically insufficient for production deployments. In this blog, we describe our war story of overcoming this gap, i.e., going from 80% to 100% accuracy in a production environment. In our experience, this journey is less about the technology itself and more about the collaboration that fuels its growth.

Will SQL be the new Assembly Language?

Accessing enterprise data shouldn’t be hard—but for billions of professionals, SQL remains a barrier. Despite decades of BI tools and dashboards promising “self-serve” analytics, business users are still blocked by complexity, slow workflows, and reliance on data experts. Text-to-SQL AI seemed like a solution, but auto-generated queries still require verification, leaving the SQL wall intact. Tursio reimagines data analytics by connecting natural language questions directly to enterprise data while ensuring correctness. By inferring semantic models, constraining AI queries to relevant data, and systematically building query plans, Tursio delivers accurate, interpretable, and actionable answers—making AI-powered analytics truly accessible.

Why AI fails despite "great” models?

We’ve all seen AI demos—type a question, get an instant answer. In reality, 95% of enterprise AI pilots fail—not because models are weak, but because prompting is hard. Most users aren’t prompt engineers, especially when querying structured data, leading to hallucinations and lost trust. At Tursio, Auto Mode solves this: it scaffolds questions based on what exists, what’s meaningful, and what’s probable in your data. Users explore, select, and accept—without worrying about SQL or prompt complexity. AI should adapt to humans, not the other way around. The result? Fast, reliable insights you can trust.

Redefining Productivity with AI

AI isn’t replacing people—it’s amplifying them. Across functions, AI acts as a force multiplier, automating repetitive work, accelerating research, and enabling faster, more confident decisions. Analysts interpret rather than wrangle data, marketers generate content and insights instantly, customer teams serve faster, and managers focus on strategy over reporting. The new productivity measure is decision velocity, not hours worked. Tomorrow’s professionals need data literacy, AI copiloting skills, and cross-functional agility. At Tursio, we empower teams to access structured data in natural language, removing barriers and unlocking human potential, so AI amplifies impact instead of just adding tools.

MCP for Databases: New trick for old elephants

In enterprise AI, making LLMs understand structured data is tricky. Many turn to LangChain for rapid prototyping, but Model Context Protocol (MCP) is emerging as a production-ready alternative—sometimes called the “USB-C for AI.” MCP servers translate natural language into schema-aware, secure queries, giving LLMs safe access to SQL, Snowflake, or FHIR databases without glue code. While LangChain excels at flexibility, MCP shines in regulated, structured environments. Tursio complements this by focusing on the human experience: effortless natural language querying, context-aware explanations, and insights for decision-makers. Together, structured access and user-centric interfaces define modern AI workflows.